Introduction

Since 2013 it has been the goal of the Dutch government to reach 100% open access to publicly funded research by 2020. 2021 will be the year in which to take stock of what has been accomplished and to look ahead to consider how to proceed with the national open access agenda and how that relates to the many developments taking place internationally, like the start of Plan S and the launch of the Horizon Europe funding framework with its progressive open access policies.

As a result of the clear target set by the government, annual progress reports to the Dutch Parliament have tended to focus on the progress that is being made in quantitative terms. Although we in no way wish to detract from the ambition to make all publicly funded research open access, we argue that there is more involved than just striving for 100%. Thus, this article is not about ways to close the gap to 100% open access. Instead, we suggest broadening the scope to include many more publication formats than the traditional research articles. But we also suggest deepening the open access properties of this broader corpus of outputs. We will present a framework that shows that open access has many more dimensions, with opportunities for action on many levels. Our argument can be summarized as a turn from a mostly quantity-oriented agenda to a more quality-oriented approach. This broadening and deepening of the agenda is not a goal in itself but serves the ultimate goals of open access: ensuring that research is being communicated and disseminated in the most effective way contributes to solving societal challenges and questions.

The scope of this article is restricted to published research outputs. Of course the aim of open science is to open up a much broader array of outputs created during the research process (data, code, software, lab notes). For the purpose of this article they are out of scope, but the approach and the framework presented could be enriched by extending it to open science practices more broadly.

Open access progress in the Netherlands

The ambition to reach 100% open access for all publicly funded research was first formulated in a Letter to Parliament by State Secretary Sander Dekker in 2013. The fact that most research is publicly funded and therefore should be openly available was an important motivation. In addition, the economic potential of open access was stressed: ‘Open access promotes the exchange and circulation of knowledge, which contributes to the innovative capacity of the Netherlands.’

In the letter, Dekker expressed a clear preference for the gold route over the green (repository) route to open access, and also he felt the need to set very clear quantitative goals. The full transition to open access should be realized within ten years (2024) with a clear intermediate target of 60% to be realized within five years (2019). These targets were revised in 2016 under the Dutch Presidency of the EU Council where the so-called ‘Amsterdam Call for Action’ expressed the ambition to reach the full transition to open access for scientific publications by 2020. This goal was then formalized in the National Plan Open Science (NPOS), published in early 2017.

The clear objective set at the national level by the Dutch government marked the beginning of an ambitious strategy from the Dutch Association of Universities (VSNU) – in collaboration with the national library consortium (UKB) – to negotiate open access agreements with the large academic publishers. The focus was first on the larger publishers. Subsequently, deals with smaller publishers were also made. Currently 18 (transformative) read and publish agreements are in place and are extended or renegotiated on a regular basis. (Note ) The agreements have contributed to a clear increase in open access availability of Dutch research output. The website openaccess.nl and the Journal Browser tool were developed to inform researchers about their open access options. (Note )

In 2015 – despite the government’s preference for the gold route – the green route to open access received important legal support. During a review of the Dutch Copyright Law the so-called ‘Taverne Amendment’ was introduced. This amendment – laid down in article 25fa of the Dutch Copyright Law – allows all researchers to share their (short) academic works through a repository after a non-specified embargo period.

In 2019 a pilot project was launched by the UKB to support authors who want to make use of that right. Guidelines were developed on what ‘a reasonable period’ would be (six months) and which version of the ‘short academic work’ could be shared (the version of record). Over the course of the year more than 600 researchers participated and more than 2,800 publications were made open access through this pilot project. Despite the fact that publishers have expressed concerns about the Taverne Amendment, no formal take-down notices have been reported, which emphasizes the value of these kinds of legal provisions in at least three ways:

  1. Providing a fallback option in cases where other routes to open access are not available yet.
  2. Facilitating open access to other formats other than journal articles.
  3. Enabling retrospective open access.

Multiple universities in the Netherlands are considering how to embed the Taverne Amendment in their institutional policies. In its 2020 letter to the Minister of Education, Culture and Science the VSNU advocated for a review of the Taverne Amendment in such a way that it would support the zero embargo sharing of papers, and would allow sharing with an open licence to make it a worthy instrument for Plan S compliance.

The ambitious government objective of reaching 100% open access received strong backing from the policies of the Dutch Research Council (NWO) which already had a long-standing commitment to make the research it funded openly available. Its first open access policy was included in its grant conditions in 2009, shortly followed by an open access incentive fund (2010) from which not only publication costs for articles and books were refunded but also projects to stimulate the transition to open access. The funding programme supported the transition of journals to open access (most notably the flip of the linguistic journal Lingua to Glossa) but also the OAPEN-NL project: one of the first studies on open access books and monographs.

In response to the government’s 100% ambition, NWO decided to revise its open access policy from a ‘strong recommendation’ into a fully fledged funder mandate. And in 2018 NWO announced that it would step up its efforts to make the research it funds openly available by joining cOAlition S, the international coalition of research funders behind Plan S. Plan S – in short – requires full, immediate open access under an open CC BY licence and with the retention of sufficient copyright for all articles that report on research funded by this growing group of research funders. NWO and ZonMw (the Dutch health research council) applied this new policy to grants arising from calls for proposals from 1 January 2021.

It is widely acknowledged that a transition to full open access can not be realized only by a combination of funder mandates and open access deals with big publishers. Open access and open science in general require a change in the reward system of science. Important steps have been taken in the Netherlands in the past few years. A note on the need to recognize and reward open science practices was written in the context of the National Programme Open Science (NPOS) which in turn inspired a major revision of the Strategic Evaluation Protocol (SEP), the national framework for research assessment. Open access and open science figure as important assessment criteria in this new protocol, which will take effect from 2021. The need to reward and acknowledge open science practices also features prominently in the position statement ‘Room for everyone’s talent’ (2019), a collective call for action by all Dutch research institutions to fundamentally rethink the system of recognition and rewards. One of the ways NWO is putting that into practice is by introducing a narrative CV format based on the principles of the San Francisco Declaration on Research Assessment (DORA).

Applicants are specifically encouraged to mention open access and open science practices. For Horizon Europe, open science practices are also expected to be taken into account in the assessment of proposals. (Note )

Meanwhile, it is an encouraging sign that the goals and practices of open access are increasingly embraced by researchers in the Netherlands. According to a 2020 survey conducted by the national organisation of graduate students, 63% of almost 1,500 respondents said they felt encouraged to publish open access. It exemplifies that open access now also profits from a bottom-up culture change.

The measures taken by the various institutions following the formulation of the government’s 100% ambition, have resulted in a sharp increase in open access availability of Dutch research articles. In 2016 – the first year a national monitor was conducted – a percentage of 42% was reported. In 2017 and 2018 this figure increased to 50% and 54% respectively. In a letter to Parliament it was reported that 61.5% of peer-reviewed articles published in 2019 by researchers affiliated with Dutch research institutions are openly available. Partly due to the agreement with Elsevier, the VSNU expects that this figure will increase to around 75% for 2020. NWO reports similar figures in its annual report. This means that although progress has been made, the official government target will not be met. A feasibility study has been commissioned by the VSNU to chart possible policy actions to be taken. Substantial extra efforts are likely to be necessary to close the gap from the projected 75% to the stated goal of 100%.

Already a number of new initiatives have started to provide or test alternative publishing routes and models. In July 2020 the open access platform for Dutch academic journals, openjournals.nl, received an NWO grant. Openjournals.nl is a stakeholder-governed journals platform to support open access journals that do not have article processing charges (APCs), i.e. diamond open access, in the humanities and social sciences. The platform will offer independent small-scale academic journals an affordable solution for publishing academic research. Another promising project, however still in its conceptual phase, is University Journals.

Despite its name, this project aims to let universities publish a wide range of research outputs, such as articles, books, protocols, registered reports and preprints using their own repository infrastructure.

All in all, 2021 will be an exciting year. The Dutch research community will be able to evaluate how successful the mix of policy interventions of the various stakeholders has been in attaining the stated goal of 100% open access. The conclusion will probably be that indeed a large share of newly published scholarly articles are openly available in some form. But new questions arise: what does it actually mean if 75% of articles are openly available? How open are those papers? Can they be reused under open licences? What about the papers that over the years have not been made open access? How should other scholarly outputs (e.g. preprints, books and monographs) be included into the open access agenda?

For a future open access agenda, we argue that there is a need to look beyond the simple rhetoric of ‘just’ attaining 100%. Closing the gap to 100% remains important but it is not the only, and arguably not even the most important, thing that needs attention now. Open access has many dimensions. We recommend the inclusion of more qualitative dimensions and to simultaneously broaden and deepen the agenda. We next turn to the question of how to implement this complex set of changes.

A multidimensional framework for furthering open access

Including quality aspects in open access ambitions requires reflection on the different aspects of open access, to address the different actors that play a role and the types of actions that can be taken. Together, these can form a framework that enables seeing where current actions are focused and where gaps and opportunities lie moving forward.

Open access is about more than just the number (or proportion) of research publications being made available through open access. Other aspects include:

  • what is being made open access
    including the sharing of preprints, other publication types such as book chapters, books, reports and older publications that could retroactively be made open access
  • how, where and when publications are made open access
    optimizing the potential of open access by removing embargo periods, enabling not just access but also reuse, making open access an available option for all authors. An additional aspect is providing rich open metadata to ensure open access publications are not only accessible and reusable, but also findable and interoperable and linked to related research outputs in a machine-readable way
  • under which copyright and licence conditions
    copyright legislation determines to a large extent what, how and when publications can be shared openly. Licences determine how reusable scholarly work can be.

Moving open access forward on all these aspects requires actions from various actors, ultimately aimed at enabling researchers to make their publications open access. At the national level, these actors include the government, national and charitable funders and academic and non-academic research institutions. At the international level, the European Commission also has an important role to play, not only as funder but also as legislator. In addition libraries, funders and universities are also organized in international networks in which they can collaborate, exchange practices or seek to align policies. We have chosen not to include publishers in our analysis here, nor researchers who as editors of (society) journals of course can play an important role in shaping open access options for their research communities. Instead, we focus here on public actors (government, funders and institutions) and the policy options they have in fostering open access.

Actions taken by these actors can range from policy setting to providing researchers with concrete support, including financial support. These actions can mutually reinforce each other to create a situation where open access publishing is not only expected, but also made possible for all researchers, and the default when sharing research results. The full range of actions open to governments, funders and institutions could be characterized as follows (Note ):

  • state as goal (make it preferred)
  • set as policy (make it required)
  • legalize and/or promote (make it known and allowed)
  • recognize and reward (make it normative)
  • finance and support of infrastructure (make it possible and easy).

It should be stressed that these levels are not hierarchical. They can reinforce each other: actions taken on each level make actions in the other levels more effective.

State as goal (make it preferred)

At the first level, open access can be stated as a goal both by governments, funders and institutions. This sends a clear message to other parties, including researchers and the general public, and can form the basis of a framework in which other actions are specified.

Set as policy (make it required)

At the second level, open access publishing can be made required, with stipulations regarding licences, limits on embargo periods and provision of sufficient metadata. This can be mandated by institutions, funders or governments for the research they finance and/or the researchers they employ.

Legalize and/or promote (make it known and allowed)

At the third level, open access publishing can be supported by creating the (legal) conditions that allow it, e.g. by provisions in copyright law that allow authors the right to share academic work financed by public money, by clauses in grant contracts that reserve authors’ right to share their accepted manuscripts and by institutional open access policies that reserve non-exclusive publication rights. Authors’ use of such clauses can further be supported by making them well-known and by providing legal back-up.

Recognize and reward (make it normative)

At the fourth level, researchers should be encouraged to make publications and other research output openly available, by making it both expected and recognized, for instance in grant applications and in career progression at their respective institutions. This includes recognizing all forms of research communication and moving away from evaluating research based on journal or publisher prestige. Instead, assessment should be based on the quality of the research itself, including openness and transparency aspects. Recognizing and rewarding publications for the quality of the reporting and the underlying research is not only a matter of policies and procedures. It is also a matter of values conveyed by researchers themselves, e.g. when they informally discuss their publications.

Finance and support of infrastructure (make it possible and easy)

At the fifth level, options to publish open access must be available, affordable and easy to use for researchers, and infrastructure must be in place (both at publishers and at institutions) to make information about publications, as well as publications themselves, findable, accessible, interoperable and reusable.

Actions can be taken on any one level by any one party – although their effectiveness does depend on actions taken at other levels and by other parties. For example, while setting goals sends a clear message, unless additional actions are taken to achieve these goals, it is likely to be ineffective on its own. Similarly, setting mandates without providing support to meet them is likely to cause resentment rather than induce behavioural change.

Ideally, open access publishing would be aligned both with the intrinsic and extrinsic motivations of researchers – motivations for sharing their research findings and for being recognized for their contributions. It is up to the various parties that shape the environment researchers work in (i.e. institutions, funders, government) to create the conditions that make this possible.

Combining the three dimensions outlined here (i.e. aspects of open access, actors involved and the levels at which actions can be taken) results in a template multidimensional framework that can inform future developments (Figure 1). Vertically, the different aspects of open access are projected, as identified in this article. Horizontally, the five ‘levels of engagement’ are presented for each of the different actors relevant for open access in the Dutch context.

Figure 1 

Template multidimensional framework for furthering open access

The framework can be used in various ways. For instance one could fill it out with current actions or policies (Figure 2) or use it to prioritize or plan future actions. Various actors could use the framework to check the consistency of their set of actions and check whether their actions align well with those of others.

Figure 2 

Framework with current open access policies and support for Dutch authors. (Note ) A full and editable version is available for download

Broadening and deepening open access dimensions

In moving beyond the goal of 100% open access for articles, we propose a strategy of broadening and deepening. Broadening includes making publications open access retrospectively in addition to looking at current output. It also includes using a more national approach including output from research-performing organisations other than universities. But here we would like to address a third dimension of broadening: making more types of outputs open access than just journal articles. Put another way: what do we make open access? Deepening refers to the characteristics of and forms in which publications are made open, or in other words: how do we provide open access? Both broadening and deepening will be facilitated but at the same time also restricted by the copyright and licensing context.

More than just journal articles

Hitherto, Dutch open access policies have been aimed at current journal articles primarily. We pose there is a need and opportunity to also look at other document types (e.g. chapters, books, conference proceedings, contributions to encyclopaedias, posters), early versions (e.g. submitted manuscripts or more generally preprints and working papers) and maybe also non-scholarly and/or non-peer-reviewed publications (e.g. reports, publications in professional and trade magazines, non-research articles, textbooks). And finally, retrospectively making any older peer-reviewed publications open access could be considered. With the exception of books, all of these have so far gotten relatively scant attention in Dutch open access policies.

Open access for books and chapters very much lags behind that of journal articles. It is of great importance as even rich libraries can only provide access to a minority of the scholarly books that are being published globally, even counting access through interlibrary loan in the Netherlands as a whole. At the same time, a study into publication cultures (Note ) showed that books and chapters represent a substantial share of the publication output, especially in humanities and social sciences. Steps taken thus far to promote open access for books are from NWO, that includes books in its Plan S-based policy and has a specific fund for open access books, and from institutions that include books in their ‘You share, we take care’ interpretation of the Taverne Amendment. Also some institutional open access funds extend to books and chapters. More systematically, there have been steps towards open access policy for books at a national level and in 2021 work will be done by UKB on approaches for monitoring open access availability of books and chapters.

In some science and engineering fields conference proceedings are very important: as important or perhaps even more important than journal articles, and open access figures for conference proceedings are much lower than those for journal articles. Also, there has been scant development of policies to increase open access for conference proceedings. (Note ) Some institutions, like Wageningen University & Research, have a specific policy requiring conference publications to be openly available.

Making early versions (preprints) open access has gained impressive momentum under the pressing need of finding solutions to the COVID-19 crisis. Sharing preprints or working papers has been a long-standing practice in fields like physics, mathematics, astronomy and economics and has also grown for a few years now in, for instance, life sciences, psychology and some other fields. There are no policies that we know of that require sharing preprints or early versions. Funders (e.g. NWO, ERC) are starting to allow mentioning preprints in grant applications and some institutions financially support specific preprint archives.

There are other perhaps less obvious document types that deserve attention in open access policies. Trade books and magazines are not primary research and have thus been out of scope of open access policies so far. Due to their often commercial nature these are slow to move to open access and any developments may depend on choices made by researchers themselves. Textbooks are another non-research type of publication that is rarely open access and deserves more attention. The idea of making textbooks open access, or rather creating them as open from the start, as online open textbooks, has spurred a whole open textbook movement. That movement is still relatively weak in the Netherlands. An exception may be the TU Delft Open Science Programme, which has strong ambitions on the support for creating and publishing interactive open textbooks and replacement of existing commercial textbooks. Another good step is the publication of an inventory of open textbooks for Dutch higher education. Finally there are the contributions to encyclopaedias, which are important for cohesion and the progress of science and scholarship and also for consolidating accepted knowledge. Currently these are rarely open access. One could consider these as short scholarly works that could also be shared under the Taverne Amendment. An alternative option would be for scholars themselves to choose to (also) contribute to encyclopaedias that are already open, like Wikipedia and topical and disciplinary ones such as the Stanford Encyclopedia of Philosophy.

Open access that is immediate, has no fees and has open metadata

When looking for options to increase the value and impact of open access, apart from applying it to more categories of academic output, we can also improve how, when and where output is made open access.

One of the main tenets of open science is increasing the engagement with research and its outputs. Having publications open access available sooner rather than later helps with that. Open access immediacy increases the chance of engagement and the possibility for verification that is especially important in the first weeks after publication, when the research results have a first spike of attention. For publications that receive a lot of media attention it is especially important for journalists and the wider public to not only be able to read the news pieces but also gain access to the publications themselves. Relying on publishers selectively giving access to science journalists is too limited. Some policies, e.g. Plan S and Horizon Europe, require immediacy of open access and UNESCO, in its First draft of the UNESCO Recommendation on Open Science, also calls for immediate open access. However, most institutional policies and mandates do not. Some of those policies prefer gold open access and with that implicitly support immediate open access, but many others rely on green open access and with that implicitly accept delays caused by publishers imposing embargoes on self archiving. The same holds for self-archived publications based on exemptions in national copyright law (like article 25fa in the Netherlands) as these often still contain embargo periods.

It is important to look at removing financial obstacles to publishing, by supporting so-called diamond open access publishing. Globally, many have criticized focussing only on APC-based journals in fostering open access. The critique is that this merely shifts the onus of payment from those wanting to read to those wanting to publish and consequently creates new inequalities. Diamond journal publishing models deserve attention to avoid the drawbacks of APC-based publishing. Existing diamond journals are sometimes regarded as not requiring any fostering as they are already fully open access and do not charge authors. However, for these (often smaller) journals to remain a viable publishing venue, they cannot be neglected. There are projects to set up diamond publishing options for institutions, such as through the University Journals project, and to support development of new and existing diamond journals in terms of infrastructure and visibility, such as through the Openjournals platform. cOAlition S also recognizes the potential importance of diamond journals and commissioned a study into the diamond landscape and the challenges for these journals. The big question is what kind of support diamond publishing venues need and how to organize that support. Coming up with solutions for infrastructural support and scalable mechanisms for financial support where necessary is also needed, and may depend on collective action, as outlined in a recent Knowledge Exchange report.

Sometimes overlapping with diamond open access, alternative publishing platforms represent a move away from the traditional journal as organizing principle. These platforms might differ from traditional scholarly journals in a number of ways, including publication process, governance and underlying infrastructure. They often apply a wider disciplinary scope, include the publication of submitted versions or preprints and apply open peer review. Often the focus is on transparency and efficiency rather than selectivity or prestige. Notable examples of alternative publishing platforms are those using the infrastructure and publishing model provided by F1000Research. But scholar-led platforms like SciPost and Liberate Science could just as well be qualified as ‘alternative’ publishing platforms.

Despite the fact that ‘alternative publishing platforms’ were mentioned as one of the five pillars of the open access strategy of the VSNU it seems these initiatives have not received the same level of attention from research institutions as has gone into negotiating transformative deals with traditional publishers. One exception, next to the aforementioned University Journals initiative, is TU Delft that aims to set up its own open access publishing platform as part of its open science programme.

And finally we need to improve the openness of metadata and infrastructure. To maximize the potential for use and reuse of research publications, not only publications themselves but also their metadata should be openly available with a licence that enables (re)use. Availability of open metadata facilitates discovery and makes it possible to create a ‘record of versions’ of a publication (including preprints and accepted manuscripts, but also updated versions) and to link publications to related research outputs such as code, data and peer-review reports. Over the past years, strides have been made to convince publishers to make citations and abstracts of publications (that can both be considered metadata) openly available (Note ), facilitating innovation and potentially lessening the dependence on commercial parties providing discovery and research evaluation systems based on closed metadata. Also, Dutch research organizations are exploring the feasibility of an Open Knowledge Base for research information, and open metadata will be an important aspect in both the development and potential use of such a resource.

A crucial dimension of open access, and something that could be given more attention in the current national open access strategy, are the copyright and licensing conditions under which publications are made available. They determine to a considerable extent whether publications can be shared through a repository and whether publications can be reused. The right to reuse is therefore considered a crucial element of the open access definition in the Berlin and Budapest declarations.

In the Netherlands, to date, only limited attention has been paid to the aspects of copyright, copyright retention and open licensing. These aspects can be included in negotiations with publishers. Additionally, actions can be taken on an institutional or funder level. A variety of strategies exist and have been tested. What follows are just a few of the possibilities that could be considered.

As is well known, many publishers ask authors to transfer (parts of) their copyright, often thereby restricting authors in sharing and reusing their work. Often the deposition in repositories is allowed, but in many cases embargoes have to be taken into account and works cannot be shared under an open licence. The most basic option available is for authors to negotiate the retention of certain rights when offered publishing contracts. Obviously this requires negotiation for which few authors are prepared. Another option is using a non-exclusive licence to publish. In this case copyright is not transferred and the rights to share and reuse are retained. A number of publishers have already adopted this practice which ensures authors retain important rights to share and make their work reusable.

Retain right to share author accepted manuscripts (AAMs) immediately and CC BY

A new strategy, which has been introduced recently by cOAlition S is the so-called Rights Retention Strategy. Its aim is to support the repository route to open access. cOAlition S funders have changed their grant conditions such that their grantees will be mandated to apply a CC BY licence to all submitted AAMs arising from their funding. Because the grant conditions predate possible later conditions set by the publisher, this allows cOAlition S funded researchers to share their work through repositories without restrictions. Recently AAAS, with their highly selective Science family of journals, has announced that it will indeed accept papers submitted under these conditions, as the only compliant route for cOAlition S-funded authors.

In a variation of this model authors are requested by their employers or funders to assign a non-exclusive licence to publish. In this case the institution or funder acts as a copyright administrator for their researchers and/or grantees. The most concrete example where this has led to an effective policy is the so-called Harvard Licensing Model, developed and first implemented at the Harvard Faculty of Arts and Sciences in 2008. Woutersen et al. have shown that this model could work in the Netherlands. This policy is especially valuable if the copyright is retained by the university’s employee.

Last but not least, the Dutch experience shows that a legally enshrined right to share can be very effective and strong. The Taverne Amendment was laid down in Dutch Copyright Law in 2015. Similar amendments are already in place or being seriously considered in other EU member states. This movement could be further supported by legislation on a European level, which indeed seems to be the intention of the European Commission. In its recent communication regarding the European Research Area (ERA) an analysis was announced of ‘authors’ rights to enable sharing of publicly funded peer-reviewed articles without restriction. To that end LIBER has published a ‘template model law’ which allows authors to retain rights and share their work without embargo or any restrictions on reuse.

How to proceed? Some concluding remarks

In this article, we argue that the Netherlands has – in a relatively short period of time – come a long way in realising the government’s ambition of 100% open access for publicly funded research. The fact that this goal was set in such a specific quantitative way by the then State Secretary Dekker has certainly contributed to this because it ensured focus in the efforts of the various stakeholders involved in the Netherlands.

At the same time, we have argued that open access has so many more dimensions than just striving for that quantitative goal of 100%. Without denying the importance of this ambition, we have argued for a more qualitative approach. An open access agenda that is both broader and deeper in scope. Broader in scope means including many more scholarly publication formats than the traditional peer-reviewed academic article on which many of the current policies in the Netherlands (and abroad) have focused so far. We have suggested a range of scholarly outputs to which the open access agenda could be extended. At the same time we argue for an extension along another dimension: the how, where and when of the openness applied to the various scholarly outputs. In its most recent progress report to the Ministry of Education, Culture and Science, VSNU has specifically extended its open access ambitions, stressing the importance of the immediacy of open access and reusability. Many more (qualitative) aspects of open access could be taken into consideration, by various actors.

In this article we have outlined what type of actions can be taken on these dimensions. Taken together this creates an ambitious – maybe a more holistic – open access agenda. To facilitate further thinking and concrete actions we proposed a framework. We do not pretend this framework is complete. And obviously it is a very time-bound exercise. The framework presents the goals, policies and ambitions at the moment of writing this article. As open access and open science are becoming the norm, the contents of this framework will change. With this framework we hope to encourage debate and discussion. How wide and how deep do the different stakeholders want to cast their net in fostering open access? How far do the actors want to go? What level of agency do they have and how can actions on different levels influence or reinforce each other?

To some this new agenda could come across as somewhat overwhelming and maybe even discouraging. That is not our intention. We would not suggest for instance that the Netherlands and all institutions must adopt this comprehensive agenda. Of course not all actions will be relevant for all organizations, not all actions demand immediate implementation and not all require a co-ordinated national approach. There may be very good reasons to concentrate on certain topics first and take on other aspects or dimensions later. But in doing that it is useful to be aware of the broader, more comprehensive agenda.

In order to help institutions to engage in a dialogue on this more qualitative, holistic open access agenda we present a framework. The framework invites discussion on how broadly each of the institutions in a given national context wants to define the open access agenda and how far each of the institutions wants to, or can, go in formulating actions in the different domains. It is clear that all stakeholders have a role to play in the research system, nationally and internationally. Universities, libraries, funding organisations and the government all have a certain agency. All of them work together somehow on a national level but in addition all these institutions are organized in international networks in which they may seek collaboration, exchange of practices or even alignment of policies. But the actions each of these organisations takes individually can influence and reinforce one another. For example, the policy goal of immediate open access as set by a funder can be supported by institutions adopting Harvard-style rights retention strategies, but could also be supported by copyright reform on a national or international level. These in turn could be supported by policy development on a supra-national (European) level.

Good co-operation at a national level between government, funder and research institutions – such as the one that exists in the Netherlands in the NPOS – is therefore important. At the same time we think the value of national co-operation should not have a limiting or slowing effect. It always remains possible for individual institutions to take a bold step on their own, to lead by example, take a certain policy action and by that promote progress among other stakeholders. Funders seem to be particularly well-placed to lead the way because they are less caught in the rat race of international rankings than other institutions.

2021 will be an exciting year for the Netherlands. A year in which the Dutch research community will take stock of how far it has come with open access. But even more it will also be a year in which we will look ahead and hopefully ask ourselves the question about how to proceed with open access. We hope that this article will contribute to that discussion but we think the presented framework can also serve as an inspiration in other countries to shape their future open access agenda in all its dimensions. We invite all institutions and other actors to consider how wide and deep they want to cast their nets in shaping their open access strategy and to consider for themselves what they are willing and able to do and with whom to collaborate.