Castells, Manuel. 1998. The Information Age: Economy, Society, and Culture—Volume III—End of Millennium. Oxford, UK: Blackwell Publishers.

Chambers, Robert. 2017. Can We Know Better? Reflections for Development. London: Practical Action Publishing.

Enghel, Florencia, and Noske-Turner, Jessica, eds. 2018. Communication in International Development: Doing Good or Looking Good? New York: Routledge.

Hess, Charlotte, and Elinor Ostrom, eds. 2007. Understanding Knowledge as a Commons: From Theory to Practice. Cambridge, MA: MIT Press.

Mansell, Robin. 2013. “Employing Crowdsourced Information Resources: Managing the Information Commons.” International Journal of the Commons 7 (2): 255–277.

United Nations Development Programme (UNDP). 1999. Human Development Report 1999. New York: Oxford University Press. http://hdr.undp.org/en/content/human-development-report-1999.

This book is about the practices of openness in international development, and how  they serve (or do not serve) inclusive development goals. It builds on and extends the theory articulated in the 2013 book Open Development: Networked Innovations in International Development, by deepening the understanding of how the design and deployment of openness in practice affects inclusion. When that book was written, open development was a relatively new phenomenon, with limited empirical research. As such, it aimed to document open development activities and their theoretical potential role in social change. Since then, a proliferation of openness practices and concomitant research have emerged, which makes this area ripe for deeper analysis.

This volume represents a synthesis of the research on open development that Canada’s International Development Research Centre (IDRC) has been supporting since 2009. Each chapter delves into different aspects of how openness functions in practice in a wide range of contexts and environments. The authors use a variety of approaches to explore the relationship between openness and inclusive development, and, in some cases, critique or problematize its connections to human development. The aim is to present empirically informed critical perspectives and constructive lessons drawn from the research to inform and improve open development practice to ensure that it fulfills its inclusive promise.

This chapter lays out the central question driving the book: what is the connection between open development practices and inclusion (in its manifold political and socioeconomic forms) in the Global South? It starts with a brief discussion of what we mean by openness, inclusion, and development, and how these concepts interconnect. It then problematizes openness and inclusive development by unpacking inequalities in the Global South—the primary research context in this book—and discusses instances when openness has led to inclusive development outcomes and when it has not. Finally, the chapter connects openness and governance of the knowledge commons, which are important to the open development puzzle overall and to the analysis in this book.

Openness, inclusion, and development are tricky concepts to unpack and to understand. The term “openness” has multiple meanings and interpretations, associated as it is with a verb (to open), a noun (openness), and an adjective (an open door). And, of late, a wide range of online content and social practices have coopted the term to give ideas, platforms, and applications a veneer of availability or transparency. Indeed, “open” is increasingly associated with nearly anything available on the Internet.

Inclusion and development are similarly contested concepts. In the contexts in which IDRC works, “to include” and “to develop” (or “Who is included?” and “What is development?”) are normative statements (and questions) that are built on assumptions about optimizing the good for the greatest number, and on goals of supporting human beings in their quest to achieve their highest aspirations. Yet unpacking inclusion and development into calculable, knowable, empirical notions is not an easy task. As goals—or rather, pillars—both seem to be a function of unique experience and positionality. For example, indicators such as gross domestic product (GDP) show something about a context in which people are living, but they do not speak to genuine happiness or well-being (which, at least for some, development and inclusion are meant to support).

You can imagine, then, that combining ill-defined concepts to describe open development might cause confusion, and, in our experience, it has. Indeed, some argue that a singular definition is undesirable and ultimately hypocritical, as it closes off other alternatives (Davies et al. 2013). Fortunately, complete agreement on a singular definition is not essential in order for learning to occur. But to help bridge at least some of these communication gaps, we have focused on a definition of “open” as it exists in practice (which is spelled out in detail in chapter 2) so there can be clarity about the actual practices under discussion. We might not all agree with the use (or not) of the word “open,” but we can point to an activity happening in the world as signifying an open practice. We have found this to be a fruitful approach.

In contrast to our attempts to achieve some clarity with the concepts of open and openness, we have allowed for flexibility around the meaning of and perspectives on development and inclusion across the contributions in this volume. As Global North authors, schooled in the social sciences in Europe and North America, we hold our own views on what constitutes development and inclusion. In this volume, we have tried our best not to impose our definitions, in order to allow the volume’s authors freedom in defining the territory according to their contexts. Thus, the meaning and expression of inclusion (who is included) and development (how and for what/whom) are determined by each of the authors of this volume, who themselves are drawing on diverse perspectives and practices in a variety of domains.

Now that we have successfully muddled the issue, let us explore what “open development” means. A few examples help set the stage and point to a tentative definition.

More children are enrolled in school in Afghanistan now than in any other time in the country’s history. Yet teachers there struggle to find materials appropriate for their classrooms that are available in the three languages taught in Afghan public schools: Dari, Pashto, and English. Openly licensed learning materials are changing this, making new and relevant content available in Afghanistan through an online portal, Darakht-e Danesh.¹ These learning resources can be adapted and adopted to fit the various curricula and contexts. They enable teachers to download, repurpose, print, copy, and use the materials—as well as to contribute their own content—and to engage students in new ways of learning.

Supporting free and transparent elections in a country transitioning from single-party rule poses many challenges. Perhaps the most critical concerns for democracy are instilling trust in the electoral process and preventing prolonged uncertainty about the results. Burkina Faso’s Commission Électorale Nationale Indépendante (CENI) confronted these challenges when administering the country’s election in November 2015. Ensuring political buy-in for the processing and publication of verified election results was one concern that CENI could address by collaborating with the Burkina Open Data Initiative² and the Open Data Institute.³ The resulting Open Elections project supported CENI with a web application that enabled citizens to access election results instantly as they were validated on Election Day. This improved the flow and accessibility of election information such that, just over 24 hours after the polls closed, results were announced, and accepted, by all candidates. The Burkina Faso experience of using open-data tools, skills, and technologies during the election illustrates the potential of sharing data to augment transparency and credibility during moments of political transition.

Rural water supplies—negatively affected by pollution and the dumping of waste and hazardous chemicals—are a crucial concern for many communities in Lebanon. And the people living near these water sources are arguably the best positioned to monitor supplies and alert municipal representatives when there are issues. So communities across Lebanon, in partnership with the American University of Beirut–Nature Conservation Center (AUB-NCC), are using crowdsourced citizen data to monitor water supplies, reduce untreated wastewater, and improve recycling and reuse methods. Village committees meet with chemists, engineers, and community development specialists from the AUB-NCC to learn how to test water supplies and exchange information on water-quality issues. Citizen scientists conduct experiments with low-cost monitoring toolkits and generate their own data on the quality of water supplies. Water committees then organize meetings to disseminate and discuss their findings with the community at large, and follow up with municipal and local legal authorities to find ways of improving water quality with new, affordable technologies. In one village, the project has even led to a permanent water-testing lab overseen by the water committee. Supported by the findings from citizen scientists and grassroot-level collaborations, the AUB-NCC is helping communities generate remedial solutions and reach out to local decision-makers, companies, investors, and other communities across Lebanon to improve water supplies.⁴

These three examples—all research case studies featured in this book—demonstrate how the open production, distribution, and use of knowledge resources can support broader human development objectives. Innovations in openness, from open government to open science, are often characterized by their potential benefit to society in many arenas of social activity and organization. Indeed, openness can affect any domain where information and knowledge contribute to social change—data, policymaking, budgets, education and educational resources, science, research data, software, and innovation, to name only a few.⁵

This book documents many examples of these activities and provides numerous cases of openness practices that contribute to achieving inclusive development aims. While open development covers a wide range of activities emerging from theoretically and practically diverse origins, at its core are questions of institutional governance in the production and sharing of knowledge resources to achieve development aims. We discuss the origin and definition of open development more thoroughly in chapter 2 of this volume.

While open development has different historical roots, the diverse practices that it implies are typically inspired by a value system that recognizes openness as a good thing. This stems from the view that open practices can offer cost-effective means of tackling pressing problems of inequity and inequality. Open educational resources (OER), for instance, offer new and virtually cost-free (to the user) resources for learning that can be adapted to local contexts, although there may be associated access costs for users, such as those for connectivity and data to access the resources. Similarly, the transparency and accountability afforded by open governance resources, like open data or open budgets, can be leveraged to uncover not only the existence of socioeconomic inequality, but also the political and economic root causes.

Openness innovations do not just potentially affect development outcomes, though. They also bring into question, rearrange, or even overturn basic assumptions about how to go about development work in practice and who should play what roles (Smith and Reilly 2013). For example, the ability to tap into collective intelligence has brought into question the predominance of the role of experts in many situations, such as when teachers take advantage of the freedom associated with OER to engage students in the creation of course content by drawing on each student’s own experiences and knowledge. Similarly, the emergence of crowdfunding and crowdlending for development-related activities like building schools or roads and supporting young entrepreneurs in rural communities challenges what was once primarily the ambit of more traditional bilateral development agencies from the Global North.

More significantly, perhaps, open development creates a space for a reboot of development theory and practice itself. Open development innovations provide novel institutional arrangements in the production and distribution of knowledge and creative works that challenge the dominant development paradigm of a market-led information society (Benkler 2006; Reilly and Smith 2013; Mansell 2013). As Mansell explains, “The ‘open’ or commons-led alternative imaginaries are characterized by some form of ‘digital resistance’ to the universal model of the information society. This usually involves some form of countervailing power, a privileging of co-operation and collaboration over competition, and innovative forms of networked collaboration often by dispersed communities” (2013, 11).

The links between openness, inclusion, and development are not always straightforward or direct. The capabilities required to access and engage in open activities vary greatly across people and populations. These differential capabilities across the world are a function of a myriad of social and individual factors, such as infrastructure for mobile phone and Internet access, affordability of pricing structures, education levels, cultural attitudes toward women and girls, and more.

To understand the research and analysis presented in this book, it is imperative to situate the work within the time and context in which it all took place (from about 2009 to 2018) and the overall trends in international development during this period. There was a major recession that affected the world’s economy (and led to disruptive austerity measures in Europe) after the financial crisis in 2008. Antigovernment protests in Tunisia in 2011 sparked a wave of political (and then repressive) action across North Africa and the Middle East, which in turn ushered in a war in Syria and the rise of the Islamic State of Iraq and Syria (ISIS), both online and on the ground. We have also seen an increase around the world in nationalism and authoritarianism and a turn away from the principles of international collaboration and cooperation. This includes movements away from earlier commitments to openness, such as Tanzania, Hungary, and Turkey withdrawing from the Open Government Partnership; the 2016 election of a US president who has undermined the open government approach of his predecessor; and the election of numerous leaders in the Global South for whom transparency, accountability, dissent, equitable development, and gender equality are not a priority, or even a part of their governance model.

These global shifts have been accompanied by the diffusion of new online communication and networking tools and platforms, particularly the increasing use of mobile technologies and social media applications. Yet the surge in user-generated content on self-styled web platforms and pages that marked earlier manifestations of the Internet have been increasingly scooped up by online data behemoths like Google and Facebook. Though the wider availability of broadband and mobile connectivity globally has provided billions of people with new communication channels and lowered historical barriers to entry in the information and communication marketplace, new users are now increasingly flocking to a narrowing number of online platforms run by only a few companies. As we see in chapter 5, in some instances, these platforms play a significant role in the potential for shared content to be discovered and used.

Indeed, the reliance on popular platforms like Facebook as the de facto digital public sphere and source for news in many countries is further limiting the possibility for genuine engagement. In contrast to the democratization of the public sphere promised by the decentralized Internet architecture, the existence of a few highly used platforms has centralized the flow of information for many individuals. This results in a high concentration of information power that can then be leveraged by new machine-learning techniques for highly targeted “precision propaganda” campaigns (Ghosh and Scott 2018, 5; Smith and Neupane 2018). This is particularly concerning in countries where ethnic or other tensions already exist. For instance, there are cases where misinformation spread through Facebook contributed to inciting violence in South Sudan, and a viral WhatsApp message provoked retaliatory lynchings in eastern India (Roose 2017). There are now daily examples in the news of how misinformation is impacting people and political systems. This threat will likely only deepen with the development of artificial intelligence techniques that allow the creation of compellingly realistic fake video and audio files of speeches (Economist 2017). This proliferation of freely shared news content as a key means of spreading misinformation and propaganda poses a legitimate threat to democracy (Persily 2017).

Along with this, the ever-increasing amount of data being generated from our online activities is having a broad and often not well understood impact on our rights online—particularly for our privacy and security. The massive increase in online activity and available user data is exponentially increasing surveillance capabilities—not only for state actors from the intelligence communities, but also for private corporations, hackers, and other rogue or non-state actors. This concurrent expansion of surveillance capabilities knows almost no state boundaries and is almost completely unregulated. Although not always well understood by the average user, these kinds of issues have a chilling effect on free expression and undermine trust in online systems over time.

Enduring issues for the bottom billion persist as well. There is, for example, worsening inequality and inequity that persist between and within countries, among rural and urban populations, between women and men, and amid historically marginalized groups, such as Indigenous and ethnic minority populations. Research from the United Nations Development Programme (UNDP) and from the World Bank shows that inequality is rising within countries, and absolute inequality, as measured by the Gini coefficient, has increased dramatically in the past forty years (UNDP 2016, 30–31).⁶ Recently, this increase has become even more dramatic, massively disadvantaging the poorest: “Since 2000, 50 percent of the increase in global wealth benefited only the wealthiest one percent of the world’s population. Conversely, the poorest 50 percent of the world’s population received only one percent of the increase” (UNDP 2016, 31).

Similarly, access to the benefits of the digital revolution remains radically uneven. Despite the rapid spread of mobile devices and broadband Internet around the world, nearly 2 billion people still do not use mobile phones, even though 7 billion people (or 95 percent of the global population) live in areas covered by a cellular network (ITU 2016). And nearly 4 billion people are not using the Internet at all, constituting 53 percent of the global population (ITU 2016). International Telecommunications Union (ITU) data show that “3.9 billion people remain cut-off from the vast resources available on the Internet, despite falling prices for ICT services” (ITU 2016). Only 15 percent of the world can access high-speed Internet (UNDP 2016, 31).

All of this means that even as broadband infrastructure, especially mobile broadband, has become more available in rural areas, last mile challenges,⁷ the cost of handsets, and access to bandwidth continue even now to confine broadband usage to wealthier segments of the population. Considering these figures in relation to the digital gender gap, women have far less access to the Internet than men in many, if not most, parts of the world. Globally, the difference is 44.9 percent for women versus 51.1 percent for men; but in developing countries, it is 37.4 percent for women versus 45 percent for men (ITU 2016, 3).⁸

In sum, open development activities are happening in a challenging climate. There is a recent slight trend where countries are actively rejecting the principles of openness, as well as an increase in the risks of engaging online, particularly in the political sphere. Genuine engaged participation among the poorest and least connected in the public sphere, both online and on the ground, remains limited. Most digital engagement still requires reliable access to the Internet, use of computers or smartphones, and expendable income to pay for data access. Sometimes these differences are referred to as “digital poverty” rather than the “digital divide.” The divide has been described as a lack of access to and full use of the information and communications allowed by the technology tools caused not just by deficient connectivity, but also by lack of demand, often related to inadequate income, and lower capacity, due to literacy or skill deficits (Barrantes 2007; also see chapter 9).

Given that the contributions of technology to social change are a function of a myriad of contextual factors, it could be that innovations in openness are contributing to and exacerbating existing inequalities. Indeed, some research on openness has uncovered examples where it has led to less, rather than more, equitable outcomes in a broad range of openness activities.

For example, in the area of open data, Rumbul (2015) looked into who engages with open government data projects seeking to increase public engagement and participation in Kenya and South Africa, and found disparities among users and nonusers along gendered and educational lines. Across the projects she examined, around 70 percent of users were male and a majority of users were educated (Rumbul 2015, 13). Indeed, both open government and open data seem to require tempered or deconstructed optimism, and both are currently suffering somewhat from the shift in global politics and political alignments that displace transparency and accountability from the main agenda. A key research finding in the Open Data Barometer (2017) is that the overall readiness of governments to adopt and implement open data initiatives has been regressing. It is also the case that government transparency in some contexts might undermine good governance (Lessig 2009; Bannister and Connolly 2011). For example, open data can be biased, contributing to misinterpretation and misuse (Zuiderwijk and Janssen 2014), and openness around more sensitive types of data can lead to violations of the right to privacy (Martin and Bonina 2013). Moreover, open government initiatives are sometimes coopted to serve a deregulation agenda that generally benefits the private sector and wealthier echelons of society. Jo Bates noted, “As the most vulnerable in society are facing substantial public spending cuts, the OGD [open government data] model risks being interpreted as, and potentially becoming, little more than a corporate subsidy” (2012).

Along similar lines, in a synthesis of citizen-voice crowdsourcing initiatives, Peixoto and Fox (2016) found only seven (out of twenty-three) initiatives resulted in a high degree of governmental responsiveness. When analyzing whose voices were being collected, they found that those most in need were usually underrepresented (Peixoto and Fox 2016, 19)—perhaps due in part to the enduring digital poverty of those communities.

Other research shows that openness practices, such as peer collaboration, are incredibly challenging in low-resource settings (Berdou 2017). In this volume, Graham and De Sabbata (chapter 5) explore the data from online peer production platforms such as Wikipedia and find that use and participation are gendered and deeply unequal geographically (see also Graham and Hogan 2013). Similarly, a study of a microwork platform in Latin America, as discussed in chapter 14 of this volume, found a skewed distribution, where a very small number of freelancers ended up getting the most jobs. Furthermore, full participation on such platforms is not possible without access to computers with the appropriate software—something that is limited in low-resource settings.

Research on the use of MOOCs found that their users tend to have both higher socioeconomic status and more formal education. One study in the United States found that users of sixty-eight MOOCs offered by Harvard University and the Massachusetts Institute of Technology (MIT) “tended to live in more-affluent and better educated neighborhoods than the average US resident” (Hansen and Reich 2015, 1245). The researchers conclude that while such digital learning innovations might raise all boats, it could also lead to an increasing divergence in educational outcomes between low–and high–socioeconomic status groups. Christensen et al. (2013) similarly found that MOOC participants, particularly in middle-income countries like Brazil, Russia, India, China, and South Africa (aka the BRICS) and in other lower-income countries as well, are more likely to be employed, highly educated, and male.

The sharing of open access scholarly journal articles can also generate adverse dynamics that further marginalize academics and research from across the Global South (see chapter 6). One mechanism for this is the increasing dependence on only a few platforms for distribution and discovery of online knowledge resources, which can lead to de facto monopolization and closure through algorithms that systematically favor some results over others. Until larger systemic issues are tackled, such as how metrics like the journal impact factor create bias by placing institutional pressure on authors to publish in particular journals (e.g., tenure predicated on publishing in top-impact journals) and journal editors’ choices of what is an acceptably relevant article to publish (i.e., those articles that will be cited), these inequalities will most likely persist despite the emergence of open access (Chan and Gray 2013; Okune et al. 2016).

The fear, then, is that openness could be reinforcing the effect of accumulative advantage where the rich get richer, also known as the “Matthew Effect” (Merton 1968). This is not a surprising development (Gurstein 2011; Toyama 2011), particularly given current levels of inequality and the rise of reactionary authoritarianism, but it is nevertheless a challenge to open development optimism and the hypothesis that it is a force for inclusive development.

On the other hand, there is research that suggests there is room for hope for the openness hypothesis. Snijder (2013) found that over 70 percent of online usage data—from a set of 137 open access academic books—came from developing countries. This provides evidence that researchers from developing countries seek out and benefit from free academic books. Garrido et al. (2016) looked at the use and perceptions of MOOCs in Colombia, the Philippines, and South Africa and found that 80 percent of users come from low- and middle-income populations, and 80 percent have only basic or intermediate information technology skills. Furthermore, they found that women are more likely than men to complete a MOOC or obtain certification. Similarly, research on a teacher-training MOOC developed by Peking University in China found that 60 percent of participants came from the less-developed regions of China (Wang et al. 2018). Furthermore, Maitland and Obeysekare (2015) found that development-focused MOOCs can contribute to building students’ social capital even beyond learning gains. And, of course, there are more examples outside of education, some of which can be found in this book.

From the research discussed in this book, we can identify three openness-related policy and practice dynamics that contribute to the unequal distribution of benefits flowing from openness in Global South contexts.

First, concerted efforts have not yet been made consistently to overcome constraints and promote effective use. Theories of (and advocacy for) openness have focused on the production and distribution (i.e., the supply) of open content, while largely ignoring, or underresourcing, the use (i.e., the demand for) of the content. The goal for many, but certainly not all, openness projects is to make digital content open, full stop—often for legitimate reasons. Furthermore, openness initiatives have mainly avoided the more contentious power relationships or have met with resistance when doing so. While we see a trend toward more ecosystemic and bottom-up approaches (some evidenced in this book), the literature produced thus far can teach us a lot about how to produce and share, but not as much about how to engage and support effective use.

Second, Global North values are embedded. Openness movements and a plurality of the openness innovations have come out of North America and Europe. In some cases, these innovations are funded by philanthropic foundations, and, in the Global South context, by international and bilateral aid agencies. This is especially true in areas such as open government data, open access, and OER. One consequence has been the emergence of openness best practices, underlying theories, and supporting technologies that implicitly contain the values and assumptions of these contexts, and that at times conflict with the realities of developing-country contexts. When this happens, these best practices can steamroll possibilities for important local adaptations.

Third, openness policy lags practice. As with many emergent technosocial innovations, it takes a while for policy to catch up. Many governments have begun to support openness at the national and international level, such as through OER policies or by joining the Open Government Partnership. Some countries, such as Uruguay, have been first movers in this space (Brazil at one time was also a first mover, but is looking less so at the end of 2019). However, the general trend is that governments have not yet figured out how to leverage the possibilities of openness to improve governance, education, health, and other development outcomes. Where there are well-intentioned policies, they are often not backed up by meaningful implementation, rendering the policies themselves irrelevant. And, in some circumstances, some open policies are short lived, undermined by policy reversals, lax implementation, underresourcing, and problems with institutionalization.

The point of this section is not to argue that openness leads directly to greater inequality—not at all. Neither is its purpose to present a wholly gloomy picture about openness activities in international development. Its point, rather, is to understand what happens in situations where openness has led to unequal outcomes, and to find alternate ways that practices and tools can be harnessed to allow openness to achieve its full potential. This is one of the aims of this book.

As is clear from this discussion, the relationship between openness, development, and inclusion is not straightforward. Open development innovations always have a form and a place, and it is the diversity of these forms and places that results in a wide range of outcomes. The chapters in this book explore a broad variety of experiences of openness in developing-country contexts, across many domains and while uncovering a multitude of outcomes. This knowledge helps to deepen our understanding of the relationship between openness practices, their contexts, and development. It is from this contextualized learning that we can move to more applicable and generalizable lessons for practice and policy. We try to draw out such lessons in the concluding chapter of this book.

Indeed, the imperative for understanding and improving practices of openness is growing. If we view openness within its historical context, it can be understood both through a range of ideas about openness (such as open government) and through theories of development in the twentieth century (such as participatory development). These historical roots are important to understand because even as openness has grown in prominence globally, so has openwashing (Thorne 2009) and counternarratives that appear to close or partition knowledge.⁹

As previously discussed, some innovations in openness afforded by the Internet—for instance, open government data and open access to scholarly publishing—are under some threat by new institutional arrangements and reactionary politics. Indeed the actual closing of open data and open government initiatives, as well as the ending of net neutrality in the United States in 2018, could signal troublesome trends for open development. Boyle’s (2003, 2008) predictions from over a decade ago (i.e., that we are headed toward a second enclosure movement rather than more openness) seem worryingly prescient thus far. We can see this happening on a technical level with the emergence of non-web-based applications and in the consolidation of online content by a handful of dominant providers. Now is the moment when policymakers and civil society need to work together to decide what our knowledge commons should and need to be, for the future, to support human potential and innovation.

So while the proliferation of open development innovation could be creating a moment of renewal for development theory and how the work of development is undertaken practically through the day-to-day work of policymaking and research, this moment will not necessarily last. The opportunity could fade if practitioners cannot harness it and make the alternative imaginary a feasible and tangible reality. This involves, in part, countering forces that are threatened by open practices and those people, places, and institutions that seek to maintain their privileged positions. In particular, the inclusivity inherent in the promise of open development must be realized to stir up and improve on the practices for which open development provides enhanced alternatives—like participatory practices or open government. Failing to leverage openness to facilitate inclusive change could translate into a closing of the alternative imaginaries that make open development so compelling.

Harnessing open development for inclusive development can be thought of as an issue of institutional governance in the production and sharing of knowledge resources to achieve equitable, inclusive development outcomes. In the same way that communities govern the knowledge commons (Frischmann et al. 2014), they also govern openness (knowledge governance is discussed in more detail in chapters 2 and 3 of this volume). This governance happens at four levels.

First, at a global level, the formal mechanisms of governance for the knowledge commons globally are embedded in relationships among and between states such as bilateral and multilateral trade-agreements. Currently, the trend is toward enclosing (i.e., commodifying) the commons through pushing for increasingly strict intellectual property (IP) regulation within global or regional trade agreements (de Beer and Oguamanam 2013).

Second, at the national level, governance happens via the state and the market. This level includes IP; telecommunications policy; the current offering of Internet and mobile services; laws around censorship and freedom of expression, surveillance and privacy, and access to knowledge (Internet rights); and various facets of culture, socioeconomics, and politics.

Third, at the provincial/state, municipal, and institutional levels, as well as via specific institutions where open development applications reside, there is often some level of autonomy in decision-making and policy direction that affects the shape and nature of openness. For example, some states and provinces might have specific open data or OER policies, and institutions such as specific universities might have an open access policy and a bespoke position on IP rights.

Fourth, there is governance at the level of the application. An application can be hyperlocal or global in scale and scope, depending on the open activity and the underlying technology that facilitates it. This includes a variety of elements that shape the possibilities and nature of human participation, such as the underlying technology and human interface; the related norms, values, and roles of the online community; and the substance and form of data, information, and knowledge. Several specific openness applications are discussed in this book, such as citizen-sourcing applications and collaborative platforms for creating and sharing OER.

The first three levels—the global, national, and state/institutional—can be thought of as the ecosystem within which open development applications, the fourth level, exist. All four levels, however, interact and affect the particular outcomes of the applications and the relative equity of their benefits.

This volume has three sections which cover defining open development, governing the open development ecosystem, and governing open development applications.

Part I, “Defining Open Development,” has two chapters that provide empirically informed theoretical underpinnings to the concept of open development. While the two chapters present different perspectives on how to frame open development, they agree that the production, distribution, and consumption of knowledge are the defining components.

Following that introductory material, the bulk of the book focuses on synthesizing lessons across research at either the ecosystem or application level. Part II, “Governing the Open Development Ecosystem,” has six chapters that examine issues and outcomes at the state and global level. These issues include the inequality of access and use of open knowledge resources, infrastructure (both technological and in terms of knowledge access), and economic and social policy.

Part III, “Governing Open Development Applications,” includes a series of in-depth treatments of interventions in specific domains of development, including governance, education, science, and informal innovation.

The concluding chapter summarizes and synthesizes the key takeaways from the book and looks forward to new issues emerging around the governance of the knowledge commons for development.

One cluster of issues that we do not cover in this book are those around openness in relation to privacy—such as issues of surveillance, data governance, and security/safety in relation to digital rights. We do have researchers from our portfolio of projects who work on these issues, but time constraints made the production of a chapter difficult. It is a rich and understudied territory, and we look forward to exploring it more in the future.

We hope that you enjoy the book.

Given that the intention of this book is both to synthesize and to draw practical lessons from a wide variety of open development activities, definitional clarity is helpful to allow some generalization and cross-domain learning. The aim of this chapter is to provide a deeper understanding of open development and present a clearer definition of the concept for use in research. In refining our understanding, the definition that we offer is derived from years of research and theory building. It provides a basis for synthesis across cases and domains explored in this book (e.g., comparing lessons from open educational resources (OER) with lessons from open government data). It also focuses more succinctly on the digital knowledge commons as a core theoretical underpinning of open development, a focus that emerged from researchers in various fields adopting this perspective. It also represents a turn away from a technology and network-centric definition.

The chapter starts with a discussion of the term open development, its evolution, and a refined definition. It provides a brief overview of key research fields, movements, and concepts that inform openness in general and open development in particular. The contributing schools of thought are useful to foreground, as they offer insights into the influences and practices of open development. We then draw on these schools of thought to refine and operationalize our definition of openness further, with a focus on openness as praxis, which is about bringing theory into action.¹ We conclude with a few thoughts on open development and its relationship to other fields of inquiry.

When the International Development Research Centre (IDRC) started exploring research and field building on open practices in different domains, the field itself was undefined. While there was an emerging interest in the potential of openness across many domains, such as data, educational practices, science, knowledge, and access to scholarly publishing, evidence for a link between openness and development was weak or nonexistent. There was also no consensus around what openness meant.

One of the earliest attempts to define open development took an inductive approach, selecting and extracting common features from a wide range of activities labeled as open. The intention was to develop an inclusive, umbrella-like definition under which the myriad of specific openness activities would fit. The resulting definition of openness was “shorthand for information-networked activities that have, relatively speaking, more information that is freely accessible and/or modifiable and more people who can actively participate and/or collaborate” (Smith, Elder, and Emdon 2011, iii). This definition highlighted two key elements: openness of content and openness to people.² While this definition included a wide variety of openness activities, it was also vague and proved tricky to operationalize from a research perspective.

This was followed by a more thorough treatment of open development, which framed openness as “networked models predicated on digital network technologies” (Smith and Reilly 2013, 3). These are models that “draw on the power of human cooperation and contain some combinations of … sharing ideas and knowledge; the ability to reuse, revise, and repurpose content; increasing transparency of processes; expanding participation; and collaborative production” (Smith and Reilly 2013, 3).

Since charting out this early territory, we have learned a lot from the experiences and struggles of the open development researchers whom IDRC has supported over the years. This experience has helped us to clarify and provide more nuance to the definition. Perhaps the greatest struggle comes from defining open. There is much debate regarding the meaning of this word, with multiple definitions being adopted across various domains and contexts. This is hardly surprising, given the relatively rapid proliferation of the use of the term and the multiple possible interpretations (Pomerantz and Peak 2016). The end result, as we detail later in this chapter, is that often these theoretical definitions of open do not match the reality of openness in practice, and the lack of a common definition inhibits the transferability of lessons learned.

Most openness research relies on definitions from the literature from within a specific domain or discipline of research (as discussed in chapter 3 of this volume). For instance, open in terms of educational resources is defined in a particular way among educators, which is different from what open means to the open government data community. Moreover, researchers referenced throughout this book found that domain-specific definitions did not always resonate with the reality of openness in Global South contexts. For example, there are cases where governments intended to share their data openly, but their efforts did not match the strict definition required by open data theorists. There are also cases where educators shared educational resources openly without a particular intellectual property license required by the widely accepted OER definition. This mismatch between definition and the reality that the researchers uncovered was a common theme emerging from many country contexts and domains of action. To exclude examples from nonconforming initiatives around the Global South would greatly limit our understanding of the reality of how openness plays out on the ground in countries where, for example, strong copyright institutions or cultures are not the norm.

Another common feature is that domain-specific definitions also tend to define open as digital content with specific features (such as an appropriate copyleft intellectual property license). We found that this emphasis distracted researchers from issues that we find to be more relevant to the overall success of an open initiative. Typically, the focus on the openness of digital content led to interventions dominated by a supply-side approach, largely ignoring and eschewing responsibility for whether the knowledge resource is part of a change process or not.

Furthermore, conceptualizing openness in terms of digital content with particular attributes makes impact evaluation challenging. Does it make sense to talk about the impact of data that is open, in terms of copyright, if it sits on a website that is never used? The upshot is that we found that researchers who adopted a content-specific definition of open struggled to assess or evaluate the impact of openness interventions. Perhaps, not surprisingly, there is little by the way of impact or outcome research in the area of open development thus far (Bentley and Chib 2016).

A key takeaway from these experiences is that digital content does not have an impact on its own; impacts emerge only through the production, sharing, and use of content. This influenced the theorizing of openness presented here that focuses not on digital content, but rather on the particular practices of producing, sharing, and using knowledge resources.

Finally, a further problem of using domain-specific definitions is that they make potentially informative comparisons across openness domains difficult, if not impossible. For example, even though they both share an open label, the definitions of OER and open government data are different, and comparing them is to compare apples and oranges. Ideally, we would like to be able to learn and share lessons, where possible, from experiences and research across domains, and, in so doing, be able to avoid making the same mistakes many times over. This will accelerate improvements in open development overall.

Drawing from these challenges and from the research found in this book, we update the definition of open development as follows: Open development is the strategic application of open production, open distribution, and/or open consumption of knowledge (often via the digital ecosystem) in the pursuit of advancing human development.

There is a lot to unpack in this definition (illustrated in figure 2.1), including what we mean by “open production,” “open distribution,” and “open consumption,” which we cover in this chapter. This definition offers several clear advantages. It provides both an abstract, crosscutting approach that can be made more specific to accommodate different openness activities and contexts. Furthermore, the greater specificity of the definition through its focus on open practices helps to address the research challenges highlighted previously.

Before unpacking the history and the iteration of the definition, we offer a few observations to help clarify and provide a contrast to past thinking.

First, the definition is agnostic as to the nature of development itself, although it tends toward a normative, or politically progressive, understanding of human development, in that the concern is about expanding human capabilities. It also generally refers to openness within international development contexts. This is consistent with earlier definitions (e.g., Reilly and Smith 2013, 32). One could imagine taking a more ideological position, whereby the definition of what constitutes development—particularly around political and socioeconomic considerations—shapes the contours of the open development definition. However, the focus here is more on the connection of openness to development and how to best engage in open activities to advance development aims, where development itself is defined by the goals of the activities.

Second, this definition places a greater focus on knowledge resources.³ The first definition of openness focused on information-networked activities, or “digital openness” stemming in part from the history of information and communication technologies for development (ICT4D) and literature on the network society (see chapter 2 in Reilly and Smith, 2013). This new definition foregrounds knowledge resources rather than the digital networks that make the production, distribution, and use of these resources possible. In so doing, this definition opens up the possibilities for integrating models of openness that are not driven by technologies. As mentioned, this change of perspective was influenced by IDRC-supported research on open development (see, for example, chapters 3, 6, 10, 12, and 13 in this volume).

Therefore, third, the definition no longer refers exclusively to digitally enabled openness. As the research over the past decade has shown, there are many, typically local, open development activities in developing countries that extend a digitally open initiative through analog means, meaning that not all open development processes are necessarily digital. For example, open budget transparency initiatives sometimes involve holding community meetings and distributing posters about government service delivery in health centers and schools. Indeed, one might argue that in some contexts, local analog (offline) transparency initiatives tend to prove more effective than large-scale national, digital open data based ones. See chapter 3 for further arguments against requiring a link between openness and digital.

At the same time, the benefits of lowered transaction costs and massively increased reach of digitally enabled (typically Internet-based) open activities have propelled the rapid increase in interest in openness since the advent of open-source software. Just as the printing press was a boon for the transmission of selective kinds of knowledge, the ease of replicating and distributing digital content through the Internet has greatly expanded the potential reach of the knowledge itself, including in analog forms.

Fourth, this definition shifts the undergirding theoretical emphasis away from (digital) information networks and the network society and grounds it in the vision of the knowledge society. This is not to suggest that networks are not critical features of today’s society, nor that seeing the world through a network society lens is not useful for analyzing open development initiatives (see, e.g., chapter 10 of this volume). Rather, this shift emphasizes a society where learning, training, and participation are core activities (Mansell and Tremblay 2013) and inclusive production, distribution, and consumption of knowledge play a key role in supporting these activities. It is the movement toward elements of social change that interest us. Thus, open development, while for the most part predicated on digital information networks, contributes to social change through innovations of open production, distribution, and consumption of knowledge resources. As de Beer writes in chapter 3 of this volume, “the tension between control over and access to knowledge is a—perhaps the—unifying thread in open development.” This has implications not just for the literature that informs open development, but for policy as well, particularly with respect to intellectual property rights (see chapter 6 of this volume).

Finally, the new definition is also much more specific about open praxis—bringing theory into action—manifesting as processes and practices. As detailed next, there are three open social processes: production, distribution, and use. The action-orientation implied in open praxis ties directly into participation and broad-based engagement, and thus into inclusion—the focus of this book. We have found that this focus on praxis, and these specific open social processes, helps to improve clarity and accuracy when engaging in research and enables comparative research through a common theoretical framing. It also makes for easier identification of other relevant literature that may refer to similar practices but use different terms. Finally, it connects openness with use and, therefore, improves understanding of potential pathways from openness to development outcomes (Smith and Seward 2017). We explore some implications of this approach in more detail later in this chapter.

To understand open development, it is helpful to understand the historical context from which the variety of new openness innovations has emerged. The working definition of open development offered in this chapter draws heavily from across these contributing schools of thought and provides an overarching theoretical framework to bring different schools of thought into conversation with each other. One hope for this approach is that it can engender a more interdisciplinary and fruitful path forward for open development, which is already a multidisciplinary field of inquiry.

These are not meant to be comprehensive discussions, but merely distillations or highlights of common recurring themes in open development literature. In many ways, these main contributing schools of thought can be seen as part of the open development literature. This discussion is also useful for pointing readers to other relevant literature that may not always be immediately associated with open development.

According to scholars, open government emerged as a recognizable term in the 1950s, largely due to journalists and newspaper editors in the United States demanding greater transparency in government decision-making after World War II.⁴ One of the first uses appeared in the foreword of a report in 1953 (Cross 1953) about people’s right to access public records and proceedings (Yu and Robinson 2012, 185).⁵ This seminal report became a foundation for journalists in the United States and the freedom of information movement more broadly. From 1953 on, open government began to resonate more widely as a principle for government transparency and accountability. In 1955, the US Congress created a subcommittee on government information where the main counsel for the committee used the term in an article on public information to convey the need for government accountability.⁶ These works on open government in the 1950s helped generate greater public scrutiny and interest in holding governments accountable and also helped to spark far-reaching legislation like the US Freedom of Information Act (FOIA) of 1966 (5 U.S.C. § 552).

Inspired in part by the digital innovation of open-source software (more on this next), open government got a reboot in the twenty-first century. This new iteration, not unlike the 1950s version, is generally undergirded by a similar idea—that democratic governments and governance processes should be open to public scrutiny, characterized by similar themes of transparency and accountability. However, the twenty-first-century version also foregrounds citizen participation and collaboration, acknowledging the importance of giving people (not just journalists) a voice in policymaking and decision-making.⁷ In some sectors, open government also serves as a platform for public-sector modernization, in order to foster greater coherence in government activities, and promote more effective public oversight of governance processes. There is also a focus on open standards and open data as a means of sparking innovation and economic growth.

Yet the relationship between open government and open data generates some debate. Perhaps it is one of the ongoing tensions within democracies, or in systems of governance that aim to be more participatory, that real transparency and accountability over high-level decision-making that really matters are often lost. Open government data initiatives can promote the perception of transparency and accountability by releasing certain kinds of public data sets, but, ultimately, if the data sets have no real bearing on power and economic relations, then substantive changes to governance (and the promises of greater transparency and accountability) are forgotten. Because discussions around open data often fill the debate space in open governance, a hazy distinction remains between the politics of open government and the technologies used for open data (Yu and Robinson 2012, 181).

Similar to how open government emerged to encourage citizen engagement with governance processes, participatory development emerged and gained momentum by trying to upend human development processes and make them more receptive to local beneficiary populations and less captured by Western donor-driven priorities.⁸ And like open government, participatory development (broadly considered) was also renewed in the twenty-first century by innovative communication networking tools.

Philosophically, theories of participatory development coalesced out of critiques of post–World War II modernization ideas that elevated Western institutions and values above the values of then-colonized, so-called underdeveloped populations. This imposition of institutional superiority only intensified with the decolonization processes in the twenty years following the war. While the modernization approach to development was a top-down, high-debt-inducing infrastructure and institutional development process, participatory development prioritized bottom-up, grassroots development processes that considered the value of unique cultural situations and environments.⁹ Couched in postcolonial, emancipatory narratives, participatory development advocates a role reversal, whereby subordinated subjects take back power from dominant decision-makers and institutions. In the 1990s, participatory approaches to decision-making in development became popular, and, over time, branched out into a broad range of practices, from participatory rural appraisal, to participatory evaluation, to participatory budgeting.¹⁰

Yet because participation, like the concept of open, has both a theory and a practice component, it can also be captured for purposes other than empowerment and transformation. The mainstreaming of participation in development, for instance, meant that simple consultations and participation in project implementation were framed in emancipatory language but were often just mechanisms for gaining approval from subjects in development initiatives planned elsewhere (Huesca 2002). What participatory development offers mainstream development now is the discursive and programmatic focus on inclusion and pro-poor development; it also indicates that the principles of development for the poorest and the marginalized remain the norm. Yet relying on broad categories like inclusion and marginalization to stand in for development that prioritizes the poor conveys a much more policy-oriented (even top-down) thinking and moves the conversation away from the individual agency inherent in participatory methodologies. Fundamentally, the idea that development should be inverted to serve not educated practitioners, but rather vulnerable recipients elsewhere, remains a perennial tension in development work.

Participatory development, like open government, was repurposed with the mass diffusion of information and communication technologies (ICTs). Mobiles and social media platforms associated with the interactivity of new iterations of the World Wide Web seemed perfect for the renewal of the idea that people needed to be at the center of their development—the “participatory potential of new connectivity” according to Chambers (2010, 28–29). And importantly, these new tools brought agency back into the development conversation. As noted by Chambers (2010, 29), “With Web 2.0 for dev, and its cornucopia of potentials through email, Internet, video conferencing, participatory GIS [geographic information system], mobile phones, SMS [short message service], blogging, Twitter and beyond, a whole new domain of participatory interaction has opened up. … The explosion of activity is based on open source technology and philosophy and participatory approaches, with continuous and multiplying volunteer contributions from within and outside. … It illustrates the runaway empowering potentials of new combinations of technology and volunteer commitment, energy and creativity. We are in a new space.”

Today, information and communication networks are continuing to change and upend how international development is conducted. About a decade ago, this was coined “development 2.0” (Quaggiotto and Wielezynski 2007; Thompson 2008; Heeks 2010a), though the extent of change continues to unfold in a multitude of positive— and negative—ways. Every new technology application designed to foster engagement and participation often brings with it a range of other, sometimes pernicious, issues. One increasing concern, for instance, is around data and the data exhaust from all the activities performed online, which affects human rights online (or digital rights), such as the right to privacy and the right to security. But the tools offered by mobile and web connectivity are still expanding possibilities for the participation envisioned by early enthusiasts for these methodologies.

The first, and perhaps most recognized, example of commons-based peer production is the open-source software production model. Open-source software is significant as a pioneering approach and inspiration for the emergence of most other open development initiatives, such as open access to scholarly publishing, OER, open government (Mizukami and Lemos 2008), and even the ICT-based participatory development methodologies highlighted earlier in this chapter. Understanding what made the open-source software production model possible is essential, therefore, to understanding open development.

Open-source software has two important, influential features. The first is the collaborative production model that leveraged the interconnectivity provided by the Internet with the knowledge, skills, and time of volunteer programmers and the sharing of source code. This new form of collaborative production enabled the development of high-quality software products by groups operating outside of public- and private-sector firms, and without the need for large infusions of capital.

The second element is the development of copyleft intellectual property licenses, which made the open-source software production model legally possible. A copyleft license uses the legalities of copyright to make it legal to freely share the software while providing, rather than restricting, user freedoms. For software, this was embodied in the following four freedoms. The first is the freedom to run the program as you wish, for any purpose. The second is the freedom to study how the program works and to change it as you like. Access to the source code is a precondition for this freedom. The third is the freedom to redistribute copies of the original program so that others can benefit from it. The fourth is the freedom to distribute copies of your modified version to others. By doing this, you can give the whole community a chance to benefit from your changes.

Together, collaborative production and copyleft intellectual property licenses formed what was later termed “commons-based peer production” (Benkler 2002). The commons-based peer production model has since evolved into a panoply of variations, including the well-known Wikipedia (Fuster Morell, Martinez, and Maldonado 2014). The movement from software to other knowledge production activities was also enabled by the development of a new copyleft licensing system via Creative Commons (CC). CC licenses provide different degrees of freedom regarding the potential use of the knowledge resource, such as to retain, reuse, revise, remix, and redistribute (Wiley 2014). These freedoms and their centrality to most interpretations of openness are discussed in more detail in the following sections.

Open innovation is another significant and influential school of thought (particularly in innovation management), with a rich and growing tradition (Huizingh 2011). There are two main subfields of open innovation: the collaborative, user-centered innovation of von Hippel and the firm-centric approach championed by Chesbrough. The democratized innovation of von Hippel (2005) focuses on the ability of firms to adopt user-centered, free innovations. These innovations are “developed by consumers at private cost,” and the innovations are not protected by intellectual property rights (von Hippel 2017, 1). This model is in contrast to the traditional innovation model, where innovations are developed by firms in a closed manner, using patents and copyrights. In this way, von Hippel’s notion of open innovation is similar to Benkler’s idea of commons-based peer production in its reliance on open resources and the ability of users to use and reuse them to fit their purposes.

By contrast, Chesbrough’s open innovation suggests that firms no longer need all requisite research and development resources in-house. This is driven by the idea that “valuable ideas can come from inside or outside the company and can go to market from inside or outside the company as well” (Chesbrough 2006, 43). While the original definition has been updated over time to reflect both pecuniary and nonpecuniary knowledge flows (West et al. 2014), the basic idea of knowledge purposefully flowing in and out of a firm is the same. The term open in this open innovation is employed to describe the firm and its relationship to knowledge resources.

One way that firms are able to attract purposive inflows of knowledge is through crowdsourcing, a term coined in 2006 (in contrast to outsourcing) to describe how firms can tap into huge pools of digitally connected cheap laborers (i.e., the crowd) from anywhere around the world (Howe 2006). Crowdsourcing involves a central entity (individual or organization) enlisting the services of any number of people through an open call, typically through an Internet-based platform (Estelles-Arolas and González-Ladrón-de-Guevara 2012). “Crowdsourcing involves participants who are invited to contribute to highly specific and predetermined tasks, whose completion requires little effort” (Berdou 2017, 19). Key to crowdsourcing is its hierarchical and centrally managed nature, in terms of both governance and technical architecture (de Rosnay and Musiani 2016).

Some definitions of crowdsourcing have also been applied to commons-based peer production models (Mansell 2013; de Rosnay and Musiani 2016). While both models involve collaborative efforts to produce knowledge resources, we follow Benkler, Shaw, and Hill (2015) and Mansell (2013) and make a distinction between crowdsourcing and commons-based peer production models. We do this for two main reasons. First, the model of governance is quite different, as commons-based peer production is not necessarily centrally managed and hierarchical. Second, they differ in their use (or not) of commons-based information resources. Crowdsourcing initiatives do not, by definition, include sharing of the resulting crowdsourced data or knowledge products. Of course, the management of a crowdsourcing initiative can choose to share the results if desired. In the international development space, the sharing of the crowdsourced data is common as part of a larger change strategy, such as for awareness raising and advocacy. See chapter 14 for examples and analysis of some crowdsourcing for development activities.

While crowdsourcing here is under the ambit of open development, other elements of Chesbrough’s highly influential conception of open innovation sit uneasily within the current open development literature. This is in large part because open in Chesbrough’s open innovation does not adopt the commons-based approach that is found at the core of most open activities. In contrast to leveraging sharing and the freedoms afforded by copyleft licenses, open innovation, as per Chesbrough, typically requires strong intellectual property rights over knowledge resources to enable firms to extract rents from subsequent innovations. How insights from Chesbrough’s firm-centric perspective might further inform open development is discussed in greater length in chapter 3 of this volume.

As with other influential schools of thought in open development, the access to knowledge (A2K) movement supports knowledge generation and sharing that prioritizes benefits to the many over the exclusive rights of a few. Although it remains highly contested territory, according to Shaver (2007, 4–6), there are essentially three basic ideas driving the A2K movement: (1) knowledge is a resource; (2) accessibility is important; and (3) governments are endowed with the means to facilitate access to knowledge and its diffusion. To explain further, this means that first, knowledge is an important resource for human well-being because it accelerates human development and innovation. In turn, the ease or difficulty of gaining access to existing knowledge is thus a significant factor in how quickly innovations can be leveraged and adopted. This includes both a concern over the cost of access, which determines how, when, and whether people can access new knowledge. It also conveys an inherent normative concern for equity, in that everyone should be able to benefit from advances in knowledge. Lastly, knowledge is public, or a common good, and thus falls under the purview of governments as the collective holder of public goods, in the context of the social compact between states and citizen/stakeholders. This suggests that governments have the political capital necessary to see through regulation and investment that support more equitable access to knowledge resources. See chapter 7 in this volume for more on the relationship between knowledge, public goods, and social policy.

The A2K movement emerged in the early years of the new millennium in response to what appeared to be the increasing enclosure of knowledge resources through Northern-led intellectual property regimes that were becoming “broader (covering more kinds of information), deeper (giving rights holders greater powers), and more punitive (imposing greater penalties on infringers)” (Kapczynski 2010, 24).¹¹ This has made intellectual property law “a central battleground in the struggles over the structure and spoils of the contemporary economy” by regulating information production strategies, appropriating value from that information in the marketplace, and also by trying to regulate everything “from how we are able to learn, think, and create together to how and whether we have access to the medicines and food that we need to live” (Kapczynski 2010, 24).

It has become so significant because knowledge resources have become increasingly important for the organization of human society in the twenty-first century—undergirding economic innovation, scientific advancement, and even ethical and human development. As such, the uneven distribution of knowledge stands to have an even larger impact on how societies function, how they cohere, how they grow, and how humans within those societies flourish (Castells 2000). Thus, knowledge and its generation are not only critical to human development but access to knowledge is grounded in basic human rights, in Article 27 of the Universal Declaration of Human Rights (United Nations 1948).¹² For this reason, one of the primary concerns of the A2K movement is to ensure that knowledge resides in the commons as a public resource, not a private one.

The A2K movement, as previously noted, highlights both the importance that knowledge plays in the development of societies and individuals and new agreements for equalizing knowledge access. For the A2K movement, openness innovations represent new ways that more flexible intellectual property contributes to knowledge being produced, circulated, and consumed to advance development. Put in other terms, openness innovations are new forms of governance of the knowledge commons.

The concept of the knowledge commons has its roots in the traditional commons as developed by Ostrom and Ostrom (1977), an interdisciplinary study of shared natural resources, such as fish stocks, water bodies, and the air (Hess and Ostrom 2007, 4). The governance arrangements of the commons exist to overcome problems thought to be intrinsic to shared natural resource commons, such as free riding and over harvesting.

The knowledge commons focuses on knowledge resources rather than natural resources. In this literature, knowledge is defined quite broadly as “all intelligible ideas, information, and data in whatever form in which it is expressed or obtained” (Hess and Ostrom 2007, 7). Through examples, Hess and Ostrom (2007, 7–8) suggest that “[k]nowledge … refers to all types of understanding gained through experience or study, whether Indigenous, scientific, scholarly, or otherwise nonacademic. It also includes creative works, such as music and the visual and theatrical arts.”

The knowledge commons, then, is “shorthand for the institutionalized community governance of the sharing and, in some cases, creation, of information, science, knowledge, data, and other types of intellectual and cultural resources” (Frischmen et al. 2014, 3). It is important to note that in this case, knowledge commons does not refer to knowledge resources, but rather to institutional arrangements (i.e., the governance) of those knowledge resources. These governance arrangements can reside at many levels, local, global, or “somewhere in between” (Hess and Ostrom 2007, 9).

It is these institutional arrangements governing the knowledge commons that allow for “overcoming various social dilemmas associated with sharing and producing information, innovation, and creative works” (Frischmann, Madison, and Strandburg 2014, 1). Note that these social dilemmas are not the same challenges for natural resource commons of free riding and over harvesting. Rather, these threats include “commodification or enclosure, pollution and degradation, and nonsustainability” (Hess and Ostrom 2007, 5). These are clearly the challenges to open innovations. Furthermore, the governance of the knowledge commons also typically comes with a normative goal: maximizing access, equity, and sustainability (Pearson and Stacey 2017).

The final contributing school of thought is ICT4D, a field that began in the middle of the 1980s as an offshoot of the information systems discipline (Heeks 2008; Walsham 2017). Information systems is not a technical field, but rather consists of the study of ICTs in society, “positioned between management studies and applied computing,” and draws on many reference disciplines (Avgerou, Ciborra, and Land 2004, 1). ICT4D furthers this multidisciplinary approach, focusing research on ICTs and social change in developing countries across a range of domains such as governance, health, education, business, and agriculture (Elder et al. 2013; Gomez 2013).

A key characteristic of the ICT4D literature is the focus on understanding how ICTs connect to socioeconomic development (Avgerou 2010; Heeks 2010b; Madon 2000; Walsham and Sahay 2006). Central to this research agenda is nuancing the relationship of technologies to social change, countering the allure of grand promises of technologically driven change to view ICTs as enablers of change embedded in social contexts and enacted through the individual agency of users (Friederici, Ojanperä, and Graham 2017; Schech 2002). Decades of studies have explored the roles and interaction of contextual factors such as levels of telecommunications diffusion, the institutional environment, socioeconomic status, and ICT skills, among others, in shaping any resultant social change and benefits related to technologies (see, e.g., Alderete 2017; Barrantes and Vargas 2016). Indeed, a corollary contribution of the ICT4D literature is illuminating the significance, role, and diversity of developing country contexts where information systems are being implemented (Avgerou 2008; Walsham 2017). In particular, this literature explores the importance of embracing both contextual sensitivity and diversity (Walsham 2001).

Open development as a potential field of research and practice emerged for some as an offshoot of the ICT4D field (see Smith and Elder 2010; Smith et al. 2011). As mentioned previously, openness was seen as a set of new social organizational models made possible with ICTs (Smith and Reilly 2013). The current definition offered in this chapter, however, is broader than these earlier definitions as it draws more heavily on the contributions and insights of the other schools of thought discussed in this chapter.

Having explored the origins of openness, we now describe in more detail its meaning within open development. In this section, we discuss the specifics of open as a process of producing, distributing, or consuming open knowledge resources (i.e., open production, open distribution, and open consumption of knowledge resources).

First, by knowledge resources, we are broadly incorporating the ideas behind the knowledge commons which includes “all intelligible ideas, information, and data in whatever form in which it is expressed or obtained” (Hess and Ostrom 2007, 7–8). As discussed previously, for the most part, the open development cases in this book represent knowledge in a digital form, although digital representation is not a necessary condition to fit within our consideration of open development.

The term open knowledge resources, then, refers to knowledge resources that are publicly shared at no cost. This content may or may not be legally in the public domain or openly licensed. This definition is in contrast to other definitions that require resources to be in the public domain or openly licensed (see, e.g., Open Knowledge International’s collaboratively developed “Open Definition”¹³). However, our definition allows us to capture the reality that much of the research in this book uncovered; for users of knowledge resources in many developing countries, the line between free, illegal, and open is often blurred at best (Smith 2014; Smith and Seward 2017). Thus, rather than excluding these activities, we choose to include them within the scope of our understanding of open development. Note that this is an analytical position rather than a normative one. We are not saying that illegally downloaded and shared resources are either good or bad, but rather that, as the research shows, the normative stance depends on both the context and one’s particular perspective.

The second definitional component is comprised of three processes of open production, open distribution, and open consumption of knowledge resources (see table 2.2). First, open production processes are knowledge production processes that take advantage of collective intelligence (Bollier 2007). In particular, the two most common knowledge production models in open development are crowdsourcing and peer production. Second, open distribution, which is also known as sharing and republishing, is the practice of making knowledge resources publicly available, which are typically (but not solely) accessed via an Internet-based platform. In other words, it is the sharing of open knowledge resources. Open distribution can have many purposes, such as contributing to the knowledge commons, achieving communications goals, or instigating behavior change, to name a few. It is useful to note that we consider transparency to be sharing with the purpose of accountability.

Third, open consumption refers to the set of uses of knowledge resources afforded by the fact that they are shared open knowledge resources. Typically, the affordances are mostly thought to be a function of different legal and technical configurations of content, but they are subject to social influences as well. One typology for open consumption practices, the 5 Rs (retain, reuse, revise, remix, and redistribution), comes out of the OER literature (Wiley 2014). Hodgkinson-Williams (2015) has extended and clarified this typology by providing a more concrete operationalization of the 5 Rs, applying it to OER (see table 2.3). For more on OER, see chapter 12.

There are a few important caveats regarding the interpretation of open that we use here. First, these processes are open when they have two characteristics: they are free (no cost) and there is no exclusion criterion (nondiscrimination) for participation in these processes (Smith and Seward 2017). These are theoretical characteristics. As the many cases in this book will illustrate, while there may be no fee charged to use openly shared knowledge resources or to engage in open production processes, there is always some cost to doing so, be it related to time or financing or connectivity costs. Similarly, there is a requisite level of access and skills, among other things, required to engage in an open practice. These issues are perhaps even more relevant in some Global South contexts, where Internet connections may be more expensive relative to income, and many may not have the requisite skills to engage. See chapter 9 for a more in-depth look at some of these factors.

Second, it is perhaps more accurate to say that the nondiscriminatory nature of some openness practices is bound to a particular community or geography. For example, HarassMap, an online crowdsourcing platform that collects incidents of sexual harassment, is restricted to inputs from Egypt. Bailur and Sharif examine HarassMap and other crowdsourcing activities in chapter 14.

Third, this definition does not cover all contemporary uses of the term open, as some have only one or none of what we call the key characteristics of no cost and nondiscrimination (see table 2.4). For example, in chapter 8, Gillwald discusses open access broadband policy. This is a form of competition policy to promote access to telecommunications infrastructure, rather than something that directly involves knowledge resources. Similarly, open access in telecommunications policy and regulation typically has two principles: price transparency and nondiscrimination. Therefore, while the price of information is made transparent, there is a fee to participate. This use of open as nondiscriminatory but with a fee to participate also can be found elsewhere. For example, open universities around the world follow the same general model: they are open to anyone to attend, but with associated student fees (Agbu et al. 2016).

In a slightly different vein, open WiFi (also known as free public WiFi; see chapter 8), offered by some municipalities as an approach to tackling digital inequality, provides nondiscriminatory, no-cost access to the Internet (a large and complex knowledge resource) (Geerdts et al. 2016). Thus, one could argue that it is a form of open distribution of the Internet as a whole, not of any specific knowledge resource.

While recognizing the potential diversity of meanings and interpretations of the terms open (openness) and development, we opted for the working definition that we have given in this chapter. Note that the definition provided in this book is intended to be neither definitive nor final. The research here highlights a diversity of meanings and interpretations of the terms open and development across contexts—a point that we attempted to incorporate into our working definition. Most of the research discussed in this volume does not employ the definition explicitly—although some chapters do (e.g., chapters 7, 8, and 9). However, the particular component practices of open (peer production, crowdsourcing, sharing, and reuse) form the main set of vocabulary that authors use when synthesizing the open development research.

Making this definition explicit helps us to more precisely delimit the contribution that this book makes to knowledge and development practice. Understanding the distinctions highlighted here allows us to understand if and how we can learn lessons across activities. It also makes it clear when open or openness has different referents, making comparisons more challenging or even impossible. This is particularly useful for those uses of open that fall slightly outside the definition focused on knowledge resources (such as the work on open broadband policy and open WiFi discussed in chapter 8, or open innovation in chapter 3).

We recognize that the definition of open development that we have offered is just one of many potential definitions. We anticipate that the definition will evolve further with time and more research. We also note that our definition is controversial for some—particularly with respect to including freely, but not legally, shared knowledge resources. For the purposes of this book, however, we believe the definition that we give here works to satisfy two important goals: first enabling comparative research that, second, respects the diversity of research contexts. In particular, the focus on openness as social praxis provides a set of practices in a framework for facilitating openness research and comparing across cases and domains (Smith and Seward 2017).

As discussed, this refined formulation of open development emerged in interaction with the research on openness in Global South contexts that IDRC supported over the last decade. Over time, we learned a lot about researching openness and applied these lessons to our definition.

One thing that we have found during the course of this research is that it is generally a better communication strategy to replace the term open with a more specific openness practice whenever possible. This specificity helps to avoid potential miscommunications based on alternative interpretations of the term open. This is one of the benefits of focusing on the specific practices of peer production, crowdsourcing, sharing, and reuse. These are more readily understandable concepts that do not carry the baggage of multiple interpretations and meanings that accompany the term open.

Second, this definition also responds to the diversity of contexts where research is taking place. Domain-specific definitions typically take on a universalistic, best practice character. We have found that these kinds of definitions sit uncomfortably with the reality of openness as it is carried out and experienced on the ground, as discussed here. Many chapters in this book attest to this disjuncture. In contrast, a focus on practices provides an abstract enough concept to capture variation across contexts without overly prescribing the attributes of the knowledge resource or specifics of the practice. For example, sharing can take many forms, and factors such as intellectual property, document format, the nature of sharing tools, and the subject of content do not determine whether you share, but rather shape the contours and outcomes of the sharing in practice. It is precisely these contours that we need to understand to promote effective sharing, rather than adhering to a predetermined, acontextual—and too often ideological—notion of what it should be. This may not be the most effective approach for advocacy, but we believe that it is useful as a critical research perspective to drive improvements and we hope, ultimately inclusive development outcomes.

Overall, the intention of this chapter is to help scope the contours and contributions of the research in this book. Definitions matter in research, and poor definitions lead to poor research. Open development research has suffered this fate at times. In the course of supporting the research in this book, we have found that the definition offered in this chapter provides a clear and highly practical framework for engaging in open development research. That said, we expect there will be those who take issue with the definition. It is our hope that, whether you agree or disagree with our perspective, you will still find illuminating and useful material in this book to take with you on your open development journey.

We know that innovation—the implementation and diffusion of new products, processes, business methods, or organizational strategies (OECD 2005, 46)—is a driver of development. Over seventy-five years of solid research proves that innovation creates jobs, generates wealth, and produces benefits for society as a whole (Arrow 1962; Marshall 1920; Rostow 1960). Innovation also has the potential to enhance individual freedom and improve people’s capability to lead longer, healthier, and happier lives.

However, ensuring that the socioeconomic benefits of innovation are not concentrated among society’s elites but rather are shared inclusively by all people is a significant challenge. The crucial importance of research on inclusive innovation has been recognized, although much work remains to be done (IDRC 2011). While the leader on inclusive innovation in recent years has been the Organisation for Economic Co-operation and Development (OECD) (OECD 2015), a number of scholars have begun to lead in the field (Altenburg 2009; Heeks, Foster, and Nugroho 2014).

Can open innovation be inclusive? Can it help unlock access to knowledge or contribute to more just, equitable, and inclusive societies? Perhaps. But, as this chapter argues, much work still must be done to align the central concept of openness applied to the phenomena of both innovation and development.

This chapter exposes and criticizes the failure of most existing research on information and communication technologies (ICTs) and open development to integrate the theories and practices of open innovation. It uses the method of meta-analysis to challenge assumptions and build an understanding of innovation across the domains of open science, open access, and open data. It brings critical insights from literature in the disciplines of law, economics, management, and public policy to bear on this problem. Finally, it considers the advantages and limitations of using the theoretical lens of open innovation to shed light on open development.

Open innovation is a useful lens for looking at open development, as the chapter concludes. Open science, open education, and open data are distinct areas where open development theories have been tested and explored. Open science, open education, and open data are also applications of open innovation models in practice. A common denominator across these various areas/applications of both open innovation and open development is the concept of access to knowledge. Open science, open education, and open data are all areas where knowledge flows are essential for innovation, and therefore development.

Among those interested in development, open innovation is often misunderstood. It is conflated with concepts like open source, copyleft, or crowdsourcing. While such concepts may be examples of open innovation practices, the core principles of open development and open innovation are not necessarily aligned. This ambiguity and misinterpretation are emblematic of broader problems with overuse of the word open, which a recent review showed is associated with dozens of different terms (Pomerantz and Peek 2016). Better understanding of the relationship between open innovation and development will help grow open development beyond its roots in ICTs, in both theory and practice.

First, in theory, anchoring the concept of open development in ICTs specifically, as opposed to innovation systems more broadly, limits its potential to explain models of openness driven by socioeconomic as well as (or instead of) technological factors. A more integrated theory has wider explanatory potential and creates greater possibilities for impact.

Second, in practice, our understanding of what open means for development is fragmented across science, education, software, data, and other domains, despite the fact that innovation happens in all those contexts. Moreover, the focus on legal and economic formality in much existing research on these topics ignores the informality of innovation practices throughout the developing world. A more crosscutting analysis offers technologically agnostic, universally applicable, and sector-neutral insights into the concept of open development.

This section of the chapter extends the theoretical framework for open development beyond its roots in ICTs, and frames openness within the broader context of innovation management strategies and policy systems. After pushing the theory of open development beyond information and communication technologies for development (ICT4D) to innovation systems, it may be possible to take the concept much further. Open development, for example, could critically influence fields such as information systems (Myers and Klein 2011). Once untethered from the sole domain of ICTs, open development could even shape mainstream development discourse in the way of other breakthroughs by authors such as Sen (1999) and Nussbaum (2011). Theoretical growth may start, however, with the far more modest move from ICTs to innovation.

Reilly and McMahon suggest that there is consensus emerging from open development research about what open means: “Open resources are defined as those which are accessible, digital, affordable, locatable, timely, sharable, and appropriately licensed. In addition, they need to be presented in a format that allows for their reuse and modification” (2015, 74). However, we know that digital resources are not accessible to many people given the levels of connectivity around the world, so conditioning open on being both accessible and digital does not work.

If access is about more than mere connectivity, then the temptation to characterize open development as being a theory formulated for ICTs, or a phenomenon driven necessarily by ICTs, creates limitations. Chapters 1 and 2 of this volume acknowledge the constraints of previous conceptualizations of open development. An early framing of open development introduced it as a set of possibilities to catalyze change through ICTs (Smith, Elder, and Emdon 2011). Smith and Reilly, leaders in the field, further suggest the idea behind open development is “harnessing the increased penetration of information and communications technologies to create new organizational forms that improve the lives of people” (2013, 4). In their foundational chapter defining open models of development, they positioned open development as an outgrowth of literature on the information society and the subsequent work on ICTs as a tool for development, so-called ICT4D (Reilly and Smith 2013). Then, in explaining the history of open models, they begin with open-source software and trace the evolution of openness from that starting point.

One possible direction is to focus on “ICT-enabled open practices” and ground theory–building in those practices (Smith 2014, sec. 1). Yet it is also possible to invert the analytical framework. ICTs may be less the tool driving social and economic reorganization than a vehicle through which social and economic changes are easily visible to community outsiders. For this, the foundational social elements, rather than technological elements, are the true drivers of information networking for innovation (de Beer and Armstrong 2015, 62).

This inverted perspective explicitly recognizes that “an ICT ecosystem is … more than just a technological system; rather, it is a social system within which ICTs are embedded” (Smith and Elder 2010, 65). Moreover, contrary to popular belief, many or perhaps most (based on Google Scholar citations) recent uses of the term open do not depend on the existence of ICTs. Digitization is not a prerequisite for openness. Offline resources and processes can be open too. Of course, ICTs may be a factor facilitating social and economic change, but ICTs are neither sufficient nor necessary for a robust and holistic theory of open development. Smith and Seward (see chapter 2 of this volume) agree.

A technologically determined foundation for open development limits its theoretical potential. I suggest there is significant value in exploring openness outside of ICT-enabled contexts. Indeed, I believe this analytical shift is crucial to understand how openness affects development through innovation generally, beyond ICT4D. Doing so allows us to see that open models are everywhere, not just online. Without disconnecting openness from ICTs, it cannot be holistically conceived, implemented, tested, and improved as a model for development. A broader, crosscutting analysis also shows that the starting point for understanding openness is not the legal terms and conditions that govern access to software, but the social norms that inspired a community to seek out the appropriate legal tools to govern cocreated knowledge.

Figures 3.1 and 3.2 demonstrate the expanded potential of reconceiving open development to cover aspects of ICTs and innovation systems, instead of a subfield or outgrowth of ICT4D research. Of course, open development might cover much more than innovation, just like innovation covers issues falling outside the concept of open development. However, enlarging the domain of open development beyond ICTs creates the possibilities of pushing the boundaries of development theory itself, as innovation systems thinking has done. Indeed, studying open development in connection with innovation systems, not just ICTs, has the added benefit of connecting the concept with a much richer body of knowledge on development generally (Lundvall 1992; Muchie, Lundvall, and Gammeltoft 2003; Kraemer-Mbula and Wamae 2010).

Furthermore, a key aspect of innovation strategy and policy that informs open development is knowledge governance. Indeed, the tension between control over and access to knowledge is a—perhaps the—unifying thread in open development (de Beer and Bannerman 2013). One of the most important legal tools for governing knowledge is intellectual property (IP). Laws, policies, and practices related to IP ownership can ultimately determine who benefits from open innovation. The potential of knowledge, in particular, to contribute to development has been well theorized and documented (Aghion, David, and Foray 2009; Juma and Yee-Cheong 2005).

There are challenges to past conceptions of openness, which both this chapter and chapter 2 confront in different ways. One challenge is that there is no common theory with which to learn and compare openness across a multiplicity of connotations, applications, and interpretations. This problem is not new. From the outset of its adoption in international development, open has been described as a fuzzy, trendy term susceptible to co-option: “a better marketing term than analytic concept” (Smith, Elder, and Emdon 2011, iii).

Davies (2012) sheds light on one reason the idea of openness is so malleable. As he explains: “There is an important distinction to observe between openness focussed on artifacts such as data, source code, or academic articles, and openness of processes, such as democracy and development.” He further comments: “Formal definitions of the former may tend to be concerned more with the legal or technical status of the artifact, whereas definitions of the latter may focus on questions of who is participating, how they are allowed to participate” (Davies 2012). Davies is among numerous researchers who suggest that openness is often defined in oppositional terms. In the legal context, for example, open is the opposite of proprietary (Reilly and Smith 2013, 31). A question to answer using a more integrated theoretical framework, therefore, is: How might the relationship between open, proprietary, and blended models of innovation enlighten us on what open development is or is not?

Another challenge, also identified in chapter 2, is that openness is typically defined in universal legal and technical terms that are not sufficient for understanding the reality of openness in a development context. While the inclusion of some legal criteria is common across the openness discourse in various domains (Smith 2014, sec. 2.2), just saying that open resources are those that are appropriately licensed is too vague to be meaningful. Indeed, there are valid concerns about whether any kind of legal licensing requirement must be part of a definition of open. “[S]coping the concept of openness as legal permissions is an arbitrary boundary that doesn’t reflect reality,” explains Smith (2016, 6–7). That is because, as explained later in this chapter, legal rules may be less influential than social norms governing the appropriation of and access to knowledge in many developing countries. Smith’s point about the arbitrariness of scoping openness through legal permissions, echoed in chapter 2, mirrors my previous point about the limitations of a technologically determined definition.

In the broader context of innovation systems, “the link between openness and innovation is not necessarily straightforward; openness may support some types of innovation, but there can be a dynamic and productive tension between open and enclosed systems. Closed technological innovations can often emerge from open systems, while open initiatives can take advantage of closed technologies or systems” (Reilly and Smith 2013, 35). Researchers have investigated whether open innovation might change innovation policymaking in catch-up economies (Karo and Kattel 2011), but they have not situated open innovation in the context of development policy more generally.

Significant uncertainty stems from the fact that many development scholars’ understanding of openness is inconsistent with the literature on open innovation, especially with respect to knowledge governance systems. This discussion now turns to the open innovation literature.

The open innovation literature is vast and well established. Numerous thorough reviews and syntheses of key sources among the thousands of works on this topic have already been published (e.g., Chesbrough and Bogers 2014; Dahlander and Gann 2010; Elmquist, Fredberg, and Ollila 2009; Fredberg et al. 2010; Huizingh 2011; Lichtenthaler 2011; Van de Vrande, Vanhaverbeke, and Gassmann 2010; West et al. 2014).

If another open anything movement is to move forward with credibility in the fields of innovation studies, management science, law, or economics, then a serious conceptual gap must be acknowledged and bridged. This is especially true with respect to open business models, where the focus is squarely on firm strategy, but the leading research is neglected (compare Mizukami and Lemos 2010 with Chesbrough 2006 and Teece 2010).

While there may be observable differences between developed and developing countries, such work should not continue in silos. It is impossible to be taken seriously by the business leaders or economic policymakers whom development researchers aim to influence without understanding the perceived connotation of terms like openness and innovation. The research gap is not just about North versus South or developed versus developing country; it is also cross-disciplinary and cross-sector. This subsection of the article summarizes the most relevant insights for open development.

“Open, user, collaborative, and related innovation concepts imply strategies and systems where ideas and knowledge flow across firm boundaries” (de Beer 2015, 11); that is the common thread in the literature on openness related to innovation. Beyond that point of convergence, however, a recent review by leading researchers identifies a “fault line” over the importance of appropriation (West et al. 2014, 808). The key divergence concerns how and why knowledge spillovers happen.

Research on user innovation, grounded in the work of Eric von Hippel, tends to view appropriation through IP rights as a drag on innovation, especially sequential or cumulative innovation, thus diminishing social welfare (von Hippel 2005). In this context, the word open is adopted as part of the phrase open collaborative innovation to mean “all information related to the innovation is a public good—non-rivalrous and non-excludable” (Baldwin and von Hippel 2011, 1400). The nonproprietary innovation is open.

Research on open innovation, a term associated with Henry Chesbrough’s seminal book of that title, sees the appropriation enabled by IP as a tool facilitating inflows and outflows of knowledge between businesses (Chesbrough 2003). Recent work on open innovation by West et al. (2014, 806) has defined the phenomenon as “a distributed innovation process based on purposively managed knowledge flows across organizational boundaries, using pecuniary and non-pecuniary mechanisms.” Chesbrough’s first book has been cited almost 12,000 times, with subsequent related works being cited thousands more times.

It may surprise open development researchers to realize that within the most widely known and cited paradigm of open innovation, strong IP protection is a key facilitator of openness. That is because the openness that Chesbrough identifies is the firm’s; in the process of innovation, the knowledge-exchanging firm, but not necessarily the knowledge, is open.

There is a parallel between the system’s/firm’s views of openness in innovation and the artifact/process distinctions of openness in development. Research following von Hippel’s work on open and collaborative innovation centers on the open nature of the artifacts circulating within an innovation system. Research following Chesbrough’s work on open innovation and open business models centers on the open nature of processes through which firms innovate.

From the firm-centric perspective, clearly delineated IP rights that can be purchased or sold by a business are among the important market institutions that explain the rise of open innovation during the second half of the twentieth century. The existing research on open innovation highlights several major trends driving the phenomenon: market institutions, labor mobility, product complexity, and technology platforms (de Beer 2015, 20).

Market institutions other than IP that promote open innovation include venture capital, securities exchanges, and other financing systems, as well as industry standards that require interoperability. Labor mobility is a crucial factor, as people no longer spend entire careers with one organization. As people move, so does knowledge. Increasing product complexity is also associated with open innovation because no single firm is able to produce all components working in isolation. Finally, technology platforms—ICTs in particular—make asynchronous collaboration across geographic boundaries possible. Technology is one (but not the only) factor responsible for open innovation, which is noteworthy for open development.

Much of open development research has seized upon only part of the literature on openness and innovation. Benkler’s concept of peer production—decentralized, collaborative, nonproprietary production by widely distributed and loosely connected individuals (Benkler 2006)—has been the most influential (IDRC 2011; Reilly and Smith 2013; Smith and Elder 2010; Smith, Elder, and Emdon 2011). Regarding appropriation (or, rather, nonproprietary) approaches to knowledge management, Benkler’s conceptual framework is more closely aligned with von Hippel’s work on user innovation and open and collaborative innovation systems than with work on open innovation in the firm-centric paradigm (de Beer 2015, 28). A key question that this chapter raises is whether widely distributed and loosely connected peer production aptly characterizes much innovation that happens in the informal sector, which dominates economic activity throughout developing countries.

Other influential concepts imported into open development research include crowd-sourcing and open sourcing, ideas that offer insights into the root causes of misunderstandings. Despite sharing the common word source, these terms cover very different ideas. Applying the various perspectives of open innovation/collaborative innovation/peer production to specific examples and applications, such as crowdsourcing and open-source software, illustrates the challenges and opportunities of integrating innovation into open development theory and practice.

Crowdsourcing is a term coined by Jeff Howe, first in an article for Wired magazine and then in a book by that title (Howe 2006; Howe 2008). A cross-disciplinary review of research later integrated forty original definitions, from over 200 sources, to propose this integrated definition: “Crowdsourcing is a type of participative online activity in which an individual, an institution, a non-profit organization, or company proposes to a group of individuals of varying knowledge, heterogeneity, and number, via a flexible open call, the voluntary undertaking of a task” (Estellés-Arolas and González-Ladrónde-Guevara 2012, 197).

With their integrated definition, Estellés-Arolas and González-Ladrón-de-Guevara (2012, 197) offer eight criteria to determine whether an activity is crowdsourcing or not:

“(a) there is a clearly defined crowd;

(b) there exists a task with a clear goal;

(c) the recompense received by the crowd is clear;

(d) the crowdsourcer is clearly identified;

(e) the compensation to be received by the crowdsourcer is clearly defined;

(f) it is an online assigned process of participative type;

(g) it uses an open call of variable extent;

(h) it uses the Internet.”

Two things are notable about the integrated definition of and criteria for crowd-sourcing. First, it is not always easy to delineate the boundaries between crowdsourcing as a form of firm-driven open innovation on one hand and open/collaborative/free peer production on the other. Take Wikipedia, for example, which in some ways seems like a form of crowdsourcing and in others seems like peer production. While there is a clear goal to create a free, publicly editable encyclopedia, via an online process involving an open (implicit and sometimes explicit) call to participate on an Internet platform owned and controlled by the nonprofit Wikimedia Foundation, there is no clearly defined recompense, at least not in the form of pecuniary compensation. Wikipedia meets more than half, but not all, of the eight characteristics of crowdsourcing. Open development researchers may conceive of Wikipedia not as crowdsourcing, but rather as a mode of peer production, in the sense imagined by Benkler. However, scholars in other fields may see it as a firm-centric open innovation strategy within Chesbrough’s open innovation framework. Ambiguities about crowdsourcing demonstrate differences between certain conceptions of peer production and open innovation, and thus raise questions for open development researchers.

Second, several of the eight elements that describe crowdsourcing seem difficult to apply in the context of informal economic activities in the developing world. In particular, much open innovation in the informal economy happens offline in social, not technological networks (de Beer and Armstrong 2015; Kraemer-Mbula and Wunsch-Vincent 2016). Moreover, the firm-centric understanding of crowdsourcing presupposes a degree of order and organization that may not be easily observable in the informal sector. This contrasts with certain kinds of crowdsourcing, such as solutions-sourcing, in the formal sector of developed countries where topics like IP management are more easily studied (de Beer et al. 2017). Thus, is crowdsourcing a misnomer in open development, or is international development missing in the research on crowdsourcing?

Open source is often understood as the inspiration for open development through ICTs. Open-source software is a better example of peer production than crowdsourcing, although many firms practicing crowdsourcing may turn to open-source solutions. The original and still-leading research explains that open source refers to software licensed on particular terms. The license must meet the criteria for free redistribution, source code availability, and derivative works, among others (de Beer 2015; Perens 1999; Raymond 1999).

One key point is that, in the context of software, open source is defined primarily in reference to legal criteria. But another key point is that open-source software is not, legally speaking, nonproprietary. It is legally protected by copyright, patents, or both, but creatively licensed to require, instead of restrict, sharing (de Beer 2015, 32). A framework is also emerging to understand open-source hardware, in addition to software, based on similar legal principles and practices.

Chapter 2 classifies crowdsourcing as a contribution that the open innovation school of thought has contributed to an overarching framework for open development. It also positions open-source software as the school of thought that contributes ideas about peer production, open licensing, and user freedom. While crowdsourcing is one example of open innovation, and software is the most dominant model for understanding openness in ICT4D, this chapter argues that looking across other domains through the lens of innovation is appropriate to frame open development. The next section of this chapter, therefore, considers the relationships among innovation, appropriation, and development in the realm of open science, open access, and open data.

Both scholarly literature and practical experience belie the premise of classifying the study of innovation, business models, or both into a silo distinct from the domains of science, education, or data. Indeed, open innovation is inseparable from open science, open education, or open data. Through access to and exchanges of knowledge, innovation is intractably embedded in, and also transcends, each of these fields. The implementation and diffusion of new products, processes, and organizational and marketing methods (i.e., innovation as commonly understood) are the phenomena that make science, education, and data drive development in a knowledge society. Figure 3.3 depicts these relationships.

Therefore, this section of the chapter considers innovation-related aspects of knowledge governance across domains ranging from science to education to data. It takes up the challenge to move “beyond studying the qualities of openness within specific localized projects, and focus its energies on crosscutting studies that identify the factors driving quality openness” (Reilly and McMahon 2015, 48). It starts with an overview of open access, which straddles open innovation and each of the three focal areas of education, data, and science most closely connected with the concept of access to knowledge as a whole, which straddles open innovation and each of the three focal areas of science/education/data.

Most definitions of openness include accessibility as a core feature. The term open access is related to, but not synonymous with, the concepts of open science, open data, and other applications of openness. It generally refers to published content that is digital, online, free of charge, and free of most copyright restrictions (Suber 2012). These freedoms address two separate barriers to accessibility. Making works available gratis removes financial barriers to access, while making works available libre removes permission barriers to access.

Both of these freedoms were emphasized in three formal declarations on open access publishing: the Budapest Open Access Initiative (Chan et al. 2002), the Bethesda Statement on Open Access Publishing (Stebbins 2003), and the Berlin Declaration on Open Access to Knowledge in the Sciences and Humanities (Max-Planek-Gesellschaft 2003). A common theme is the understanding that: “[f]or a work to be OA [open access], the copyright holder must consent in advance to let users ‘copy, use, distribute, transmit and display the work publicly and to make and distribute derivative works, in any digital medium for any responsible purpose, subject to proper attribution of authorship’” (Suber 2012, 8).

More so than open science, the term open access has long been associated with IP licensing, particularly copyrights. Imbued with this legal connotation, open access has since evolved in three related but distinct directions. First, the original issue of access to published scholarly literature continues to generate debate. Second, recent innovations have led to an increased use of open access to publish other types of content. Third, open access is moving to educational resources, both in artifacts and the process of learning.

In the first area, open access to scholarly literature, one of the strongest arguments for open access to literature is its potential to increase the impact of scholarly publications (Hess and Ostrom 2003). Early research suggested that journal articles received more citations when openly published (Antelman 2004; Eysenbach 2006). These findings generated significant interest within academia, leading to the “green and golden roads to open access” (Harnad et al. 2004, 310). By 2004, 90 percent of academic journals surveyed in one study were “green”—that is, they allowed authors to self-archive—while 5 percent of academic journals were “gold,” which means publishing under open access licenses (Harnad et al. 2004, 313). More critical reviews call these impact effects into question, pointing to variability in findings and identifying various biases (Craig et al. 2007). However, a strong consensus remains that open access increases academic impact (Tennant et al. 2016).

The demand for open access to literature has never been higher. Elsevier’s recent legal victory drew attention to the academic piracy site Sci-Hub, whose users downloaded 28 million papers between September 2015 and March 2016 (Bohannon 2016). However, limited research examines the economic impacts of open access to literature. A report to the United Kingdom’s Open Access Implementation Group estimated that open access saves the public sector £26 million in access fees and £2.6 million in time (Look and Marsh 2012). Recognizing the economic and social value of open access literature, the Barack Obama administration announced in 2013 that federally funded research in the United States would be freely available within one year of publication (Stebbins 2013).

Second, as Internet bandwidth has increased, so have “remix culture” and the demand for freely available content (Lessig 2004, 965). The Creative Commons (CC) suite of copyright licenses has helped to meet this demand, allowing creators to publish their works gratis while selecting a range of libre-based requirements for attribution and use.¹ Over 1.4 billion works are openly accessible under CC licenses (Merkley 2018). Major Internet services offer options to publish under CC, including content on Wikipedia, video on YouTube, music on Soundcloud, and images on Flickr.

Third, open access to educational resources has moved beyond textbooks and other content-based artifacts of education, to the process of learning. Educational resources are made open with the understanding that “access to the world’s knowledge is a public good” (Smith and Casserly 2006, 10). Research in the field explores how open educational resources (OER) are produced, delivered (Wiley, Bliss, and McEwen 2014), and adopted, as well as their subsequent impacts (also see chapter 12 for more on OER). OER include open courses and learning materials, software for learning (e.g., learning management platforms), and repositories of learning materials (Downes 2007).

Downes (2007) highlights a number of business models being used to provide OER. Models include an endowment model, where funding comes from interest generated from large donations (e.g., Stanford’s Encyclopedia of Philosophy); an institutional model, where funding is part of the organization’s operational budget [e.g., the Open Courseware program at the Massachusetts Institute of Technology (MIT)]; and a conversion model, where for-profit organizations offer freemium options, with the goal of converting free consumers into paying customers (e.g., Coursera) (Downes 2007). Some governments have shifted spending from proprietary to open resources.

In terms of access to the process of education, massive open online courses (MOOCs) have emerged as particularly interesting (Haggard et al. 2013). These online courses are open for anyone to enroll in and access reading lists, lectures, and various learning experiences. Examples include edX, run by a consortium of universities including Harvard, the University of California, Berkeley, and the University of Adelaide; Coursera, run by Stanford, Princeton, and Arizona State University; and Udemy, a nonacademic skills training site that allows experts to create and administer courses.

As with OER, MOOCs have struggled to find sustainable business models (Wiley, Bliss, and McEwen 2014), although there are some emerging models such as charging for specializations, on-demand MOOCs, and verified certificates (Tirthali 2016). What is clear is that the development and maintenance of online learning materials are costly. For example, each course published on MIT’s Open Courseware takes a minimum of 100 hours of effort to produce, at an annual cost of US$3.5 million.² The costs to produce and deliver MOOCs can also be quite high. A 2014 study found that the cost per course for MOOCs ranged between US$38,980 and US$325,330 (Hollands and Tirthali 2014), and a more recent paper noted that there are examples of high-end courses costing up to an estimated US$1 million (Tirthali 2016). These costs highlight the challenges that providers face in covering their costs while maintaining free access to course materials.

A number of other governance challenges exist for MOOCs. Although free, as in gratis, remains a common theme across MOOCs, now many courses are not offered free, as in libre. The debate about whether these courses are massive, open, and online or just massive and online demonstrates the ambiguity and controversy about what open really means. These developments have led some commentators to speculate whether massive online courses can remain free and sustainable (Tirthali 2016; Gee 2016).

Open access to both educational resources and processes is especially important for developing countries, as IDRC-supported research on open development clearly demonstrated (Armstrong et al. 2010). In the knowledge economy, access to information drives innovation and growth. Limited financial resources within developing countries constrain access to proprietary journals and other creative resources that are necessary for innovation. Education for innovation requires access to both materials and pedagogical tools for learning. As emerging concepts that require innovation in their implementation, open access and OER will also create opportunities for education in the process of innovation.

Moving forward, a leading thinker on OER forcefully argued that there is no longer any debate about the meaning of open in the context of OER: “As far as I can tell, the only people actively engaged in a debate about the meaning of the word ‘open’ in the educational context are (1) those who genuinely misunderstand it because they haven’t become part of the community yet, and (2) those whose business models would collapse if the public had free access to and open licenses for their products” (Wiley 2016). The key phrase in Wiley’s comment is “in the educational context.”

Cutting across sectors—looking at innovation in education, science, data, and other areas—there is significant debate about what open really means. When organizations around the world say that they practice open innovation, it is not just naivete, or “openwashing” as Wiley suggests (Wiley 2016). These organizations may be adopting a bona fide understanding of openness based on renowned work by authors such as Henry Chesbrough. Work in the realm of open data helps to further highlight the conceptual challenges that still plague open development across domains.

Data have become valuable with the rise of the digital era, informing and driving scientific discovery, underpinning business models, and supporting evidence-based policy. Developments in linked data provide incentives for firms to open their data for collaboration.

The Open Data Handbook (n.d.) defines “data” as data sets as opposed to a single data point, which it defines as “content”. Data are considered to be “open” when they can be accessed, used, modified, or shared by anyone (Open Data Handbook n.d.). Within the open innovation literature, there is some consensus that open data require ICT—the data must be machine readable and accessible online (Chan 2013; Gurstein 2011). In contrast, open development scholars acknowledge that data take many qualitative forms, including maps, pictures, and paintings, and need not be machine readable or accessible online to be open (Hossain, Dwivedi, and Rana 2015; van Schalkwyk et al. 2014).

As an intangible resource, IP governs the ownership of data. By default, owners hold a proprietary right to their data. This is a result of a combination of legal mechanisms governing ownership of data, including copyright, sui generis database rights, technological protection measures (TPMs), trade secrets, and privacy (de Beer 2016). Copyright (the most commonly discussed mechanism) automatically protects original compilations of data in databases, but not fact-based content like statistics or formulas. More than one of these legal mechanisms may apply to a given data set, increasing the ambiguity of data ownership. Therefore, data are only considered “open” when they are published under an open license.

Proprietary ownership of data poses several challenges for open data artifacts. In practice, proprietary data hinder the growth of the linked data commons, making it impossible to apply open licenses to linked data sets containing proprietary data. The ambiguous nature of ownership rights in data means that licensing open data may not be as straightforward as other content. Conversely, clearer ownership rights could facilitate open data by enabling owners to apply open licenses to their data. Open data licenses transfer these rights, allowing anyone to access and use the data with attribution and other optional conditions (Open Data Commons n.d.; Creative Commons 2013).

Despite the challenges, firms are practicing open innovation by releasing their private data, recognizing the economic benefits and potential of open data systems (Dodds et al. 2014; Hammell et al. 2012; Open Data Institute 2016). Depending on the particular model adopted to share open data, these practices could reflect either the firm-centric version of open innovation, by using IP licenses to exchange data into and out of the organization, or a more systemic approach, in which data are put into the public domain without any IP protection at all. In either case, the economic impacts of open data are substantial.

It has been suggested that the market impact of public-sector information across the European Union (EU) in 2008 was €28 billion (Vickery 2011). An independent review in the United Kingdom estimated benefits of public-sector information at £6.8 billion in 2013 (Shakespeare 2013). Looking globally across public- and private-sector information, the McKinsey Institute estimates the widespread use of open data will globally unlock $3.2 trillion in economic value per year (Manyika et al. 2013). However, it is difficult to assess the reliability of these figures because little or no empirical work has been conducted on the macroeconomic impacts of open data.

Using business case studies, the Open Data Institute (2016) describes how three firms, Thomson Reuters, Arup, and Syngenta, have adopted an open data approach to gain an economic and competitive advantage. Thomson Reuters opened access to Permanent Identification (PermID), its key entities identifier system, to improve the richness and accuracy of their data (Dodds et al. 2014). They enjoy a reciprocal benefit when outside users link their open data to the PermID system, gaining access to outside data with little additional effort (Dodds et al. 2014). Arup, a design and engineering consulting firm, uses open data to enable nimble responses to new ideas inside and outside the organization. It allows Arup to “create IP without having to have complex legal agreements, lawyers, and background discussion that slows everything down” (Open Data Institute 2016). Syngenta, a global agricultural firm, has published six open data sets as part of their Good Growth Plan for sustainably growing their business by improving global food security (Open Data Institute 2016). Beginning as a strategy to engage stakeholders and build trust in their plan through transparency, Syngenta AG view open data as part of a shift toward a more collaborative business model.

There is considerable optimism surrounding open data among the development community. The World Bank sees their potential to help level the playing field for the communication of knowledge (Walji 2011). Such inclusivity is expected to generate more effective outcomes (Davies and Edwards 2012). For example, increased access to information allows policymakers and aid funding agencies to make evidence-based decisions (Linders 2013). But scholars warn that “openness must serve the interests of marginalised and poor people” (Davies and Edwards 2012), which in many cases is overlooked and has only a tangential bearing on the open data work being done.

Several challenges must be overcome before these opportunities are realized. Unequal access to ICT and open data creates a digital divide, excluding marginalized users from the benefits of open data while empowering users who already have expertise and access (Gurstein 2011). Developing countries and businesses often lack the capacity to digitize their data (Davies 2014).

When open data are available, they are often published in the aggregate (for a variety of reasons, including privacy). But the developing world needs disaggregated data to make evidence-based policy decisions (Addison et al. 2016; Chan 2013). Data intermediaries are recognized as part of the solution to these challenges (van Schalkwyk et al. 2014), but more work is needed to increase ICT penetration and address the context of developing countries. (See chapters 10 and 11 in this volume for more on open government data.)

Science is perhaps the broadest domain for open development research because it both produces and depends on scholarly literature and data. In a way, open science subsumes and transcends many aspects of open access publishing and open data (although, as chapter 13 in this volume suggests, open science may be interpreted differently in the context of development). It may be reasonable to treat certain open access issues, such as access to scholarly literature, primarily under the rubric of open science. The framework offered in this chapter, however, considers scholarly publishing alongside other forms of open access that share common roots (specifically, access to remixed cultural content and access to educational resources). Open science has different and earlier historical roots.

Dasgupta and David (1994, 487) explain the “new economics of science” as a synthesis of “the classic approach of Arrow and Nelson in examining the implications of the characteristics of information for allocative efficiency in research activities, on the one hand, with the functionalist analysis of institutional structures, reward systems, and behavioral norms of ‘open science’ communities—associated with the sociology of science in the tradition of Merton—on the other.” The stylized term inspired by Merton (1973), open science, is now understood as being an institution with a self-reinforcing code of conduct. The scientific priority of original discovery is rewarded with monetary and social benefits, which create incentives for full disclosure and diffusion of scientific knowledge.

The open science model contrasts with one where exclusivity of property rights is the key incentive for investments in science. The tension over openness was described by Paul David and Bronwyn Hall (2006, 767) in this way, in the introduction to a special issue of Research Policy on IP issues affecting science: “[W]e have two distinctive regimes or environments for the conduct of research: the actors in the realm of ‘open science research’ expect reciprocal sharing of discoveries among themselves and the rest of the world, while those in the world of private profit-oriented and proprietary R&D expect to receive payment for the right to use their inventions (and to pay others for the use of theirs).” This and other seminal literature on open science define openness at least partly by reference to its opposite, proprietary science (see also David 2004). For more details, see chapter 13 in this volume.

The incentives produced by open and proprietary approaches have different impacts, depending on the system in which science operates. Indeed, it has been shown that open science contributes to scientific and economic inequalities in developed and developing countries (Carillo and Papagni 2014; also see chapter 13). That is because the larger the scientific system in terms of investments and involvement of the state in which one operates, the greater the social and professional rewards for openness tend to be. There are fewer rewards in smaller systems characterized by scant investments. Consequently, economic modeling shows a world with two stable equilibria, which “may explain the huge differences existing between scientific sectors of less developed and more developed countries” (Carillo and Papagni 2014, 52).

Distinctions between open and proprietary models of science had roughly tracked the divide between university- and industry-led processes. But this line is becoming increasingly blurry as universities face pressure to commercialize scientific research and industry experiments with new appropriation strategies that support more open models of innovation. IP rights play a central role here. One of the key conclusions flowing from leading research on academic science and entrepreneurship is that IP rights such as patents are “changing the ‘rules of the game’ for scientific exchange, scientific credit, and the commercialization opportunities arising from scientific discovery” (Jaffe et al. 2007, 575).

The most significant shift in academic entrepreneurship during the twentieth century occurred as a result of legislation in the United States, the Bayh-Dole Act, which permitted recipients of public funding to obtain private IP rights over research outputs. While the legislation, passed in 1980, may not have fundamentally changed processes at the few American universities that were already actively exploiting IP (Mowery et al. 2001), for most institutions the Bayh-Dole Act was “a major impetus towards increased university involvement in patenting and licensing” (Sampat 2006, 781; also see Berman 2008, 836–837). Indeed, Bayh-Dole is associated not just with increased IP awareness and acquisition, but with academic entrepreneurship more generally (Aldridge and Audretsch 2011; Grimaldi et al. 2011).

“[T]he intellectual foundations for this sea-change in federal patent policy were weak,” explains Sampat (2006, 773), “based on a lack of understanding of the roles of universities in the innovation system.” Nevertheless, in the early twenty-first century, the impacts of the Bayh-Dole Act, combined with the World Trade Organization’s Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), have been rippling throughout the developing world. In a must-read review of economic evidence pertaining to IP’s impact on science and technology in developing countries, Forero-Pineda (2006, 810) summarizes: “Scientific communities in developing countries are particularly vulnerable to limitations of cooperation and access to information, resulting from stronger intellectual property rights protection, as their efforts to obtain normal science results must be considerable.”

There are basically two ways in which IP threatens open science: access costs and transaction costs (David 2004). Institutional innovations that privatize and commodify knowledge “have a potential to do serious damage in the field of scientific and technological research, with all the adverse implications that this may carry for the long-term course of innovation and economic welfare growth in the advanced, ‘knowledge-driven’ economies and the developing economies alike” (David 2004, 10). Empirical studies, using methods such as difference-in-differences estimators, show substantial benefits of academic openness for innovation (Murray et al. 2016). Economic modelling also shows that under the right conditions, open science is associated with higher social welfare than secrecy (Mukherjee and Stern 2009).

In developed countries, a possible “tragedy of the anti-commons” is among the biggest concerns around the push to propertize and commercialize scientific research (Heller and Eisenberg 1998). Empirical research is scarce, but some evidence shows that patents do have a “modest anti-commons effect” that hinders the free flow of scientific knowledge (Murray and Stern 2007, 649). A core challenge is developing an empirical methodological framework to test anti-commons and open science theories. This is very difficult in developed countries with data about patents and publications (Murray and Stern 2007, 656–661). It is even more difficult to conduct quantitative or statistical analysis in developing countries, where new concepts and metrics to measure informal innovation are just beginning to emerge (de Beer, Fu, and Wunsch-Vincent 2013). Numerous researchers have called for reform of the Bayh-Dole system in the United States (Kenney and Patton 2009; Rai and Eisenberg 2003). There is also significant debate on whether the Bayh-Dole approach is appropriate for other developed countries. The most credible research suggests not: “efforts at ‘emulation’ of the Bayh-Dole policy elsewhere in the OECD are likely to have modest success at best without greater attention to the underlying structural differences among the higher education systems of these nations” (Mowery and Sampat 2005, 117). The problem is even worse when transplanting Bayh-Dole to developing countries (So et al. 2008, 262).

It is not surprising, therefore, that the emerging research on open science in development characterizes the issues very differently. Development-related research on open science frames the issues less in economic terms than as about “the ways in which colonial legacies, capitalist forces, and political repression continue to limit access to knowledge around the world” (Albornoz 2016, sec. 1).

While it is justifiably argued that “Open Science addresses development by expanding opportunities to create, share, and use knowledge” (Albornoz 2016), these are unfortunately not the terms on which policy debates are unfolding in real time in developing countries. New research by the Open African Innovation Research (Open AIR) network demonstrates that open science is under threat as a result of Bayh-Dole–like initiatives in numerous African countries. The conclusion from Open AIR’s work in three countries thus far is as follows (de Beer et al. 2014, 391):

It remains to be seen whether such an orientation, fashioned more than three decades ago in the world’s strongest economy, will be helpful in contemporary or future African contexts. The evidence provided in this book suggests that the IP commercialisation orientation for public research outputs will have a relatively benign impact in South Africa; potentially damaging consequences in the context of Ethiopia (with its moribund university–industry linkages); and highly uncertain results in Botswana (where the policy-making is very recent and awareness among public researchers very low).

The main reason such a dramatic policy shift away from open science may not have serious unintended consequences in South Africa is that relatively sophisticated stake-holders are crafting workarounds to avoid the worst outcomes (Ncube, Abrahams, and Akinsanmi 2013). Not all researchers are optimistic that problems can be avoided, however (Barratt 2010). Regardless, the South African approach sets a dangerous precedent for the rest of the continent, as well as the developing world more generally. Concerns have been expressed about the threats from the propertization of science in research focusing on other developing countries and regions, including India (Sampat 2009) and Latin America (Forero-Pineda 2006). Future open development-related research on open science cannot ignore this crucial issue.

“The case for open scientific knowledge clearly needs to be reconstructed,” argues at least one leading innovation economist, to acknowledge that “the scientific commons is in danger, the costs of having it erode further are likely to be high, and that we ought to move to protect it” (Nelson 2004, 456 and 470). Much work remains to be done to understand when and where an open or proprietary approach makes more sense, and time is of the essence.

Interestingly for the conceptual framing of this chapter, ICTs factor very little in the leading research on open science. There are two possible inferences to draw from its absence. Either there is a major gap in the research, with open science scholars failing to recognize the central role of ICTs and give them due analytical attention, or ICTs may not be the definitive driver of open science. If the latter explanation is true (or even possible), open development researchers would do well to study the broader social and economic forces at play when building grounded theories about openness in science, technology, and innovation policies for development.

Another interesting contrast exists between research on open science and open innovation regarding the direction of the shift toward openness. In the realm of science, openness is being supplanted by appropriation through IP rights as part of the push toward the commercialization of publicly funded research. As a result, science is under threat of becoming less open. In the realm of innovation more broadly, industry is becoming more open. Old models of closed innovation premised on isolation and secrecy are being replaced by more collaborative models based on the flow of knowledge across firm boundaries.

Is the trend similar with respect to other areas of open development, such as education and data? The following section of this chapter explores that question.

This chapter has highlighted theoretical gaps between the fields of open development and open innovation. It has suggested strategies to better integrate these two concepts, thus extending the potential of open development beyond its origins in ICT4D.

Significant discussion of the meaning of openness was contained in both major sections of the chapter. That discussion reveals that critics of openwashing may misunderstand the reason that companies call practices open, despite not meeting the criteria befitting one particular application of openness or another. The reason is that some practices may be labeled open according to the widely accepted meaning of openness in the management and innovation literature. This is not always a disingenuous or nefarious marketing ploy; it is an innovation strategy taught in introductory-level business school courses.

This chapter has also shown how the concepts of open innovation and access to knowledge provide crosscutting analysis of commonalities among applications including open science, open education, and open data. The practices that underpin open science are, in fact, modes of open innovation. Open education is itself an open innovation, and, whether they realize it or not, all (and maybe most) actors in the open education ecosystem are open innovators. Open data are among the key currencies exchanged via open innovation. Open innovation does not belong in a category with open science, open education, and open data; it simultaneously subsumes, underpins, and permeates such things. A new, theoretically integrated, and practically cross-cutting framework, as presented in figure 3.3, can facilitate more interdisciplinary and policy-relevant research.

As a final remark, there are two areas that stand out as priorities for future researchers who might use these insights to advance open development research. One relates to informality, the other to governance.

Regarding the informality of openness, it is critical to note the size and importance of the informal economy in many developing countries. The latest figures show that the informal economy contributes nearly two-thirds of gross domestic product (GDP) in the region of sub-Saharan Africa, and, in some Southeast Asian countries, such as India, informal economic contributions constitute half of GDP. Groundwork has recently been done to prove that (1) extensive innovation happens in the informal economy (de Beer, Fu, and Wunsch-Vincent 2016); and (2) informal-sector innovators openly collaborate or appropriate differently from their counterparts in the formal sector (de Beer and Wunsch-Vincent 2016). An agenda to better measure and value these contributions has been proposed, but the hardest work of developing and implementing new metrics remains to be done.

Despite the importance of the informal economy in developing countries, little or no research has been published that explores open science, open education, or open data in this specific context. An opportunity exists for researchers to take up this challenge by situating those topics in a framework integrating open innovation and development. Without more explicit and systematic coverage of the informal sector, open development cannot fulfill its explanatory potential or have the impact that it otherwise might.

On the issue of governance, further research is needed to understand the best approaches to mediate tensions among openness, access, and the appropriation of knowledge. Knowledge flows throughout innovation systems and influences the degree to which the benefits of science, education, and data lead to positive development outcomes for all of society. Access to knowledge was identified as a central concept animating the last decade of work on ICT4D (de Beer and Bannerman 2013). Now the time is ripe to take an analogous step in the emerging field of open development. It is heartening to see Smith and Seward accept that challenge, as they did in chapter 2.

One conceptual tool to better understand how knowledge governance affects innovation across science, education, data, and other domains is the commons. Numerous scholars have been inspired by the work of Nobel laureate Elinor Ostrom on governing the commons (Ostrom 1990). Her insights have been adapted to create a framework for research on governing the knowledge commons (Hess and Ostrom 2006; Frischmann, Madison, and Strandburg 2014). Interesting scholarship using this framework has been published on commons-driven initiatives in genomics, medical research, astronomy, aviation, journalism, military technology, and much more. Chapter 2, picking up on this idea, identifies the potential of a knowledge governance framework in open development. But the knowledge commons framework for governing open innovation in developing countries has not yet been well developed, nor well studied.

Informal innovation and knowledge governance are just two specific priorities that might be possible to explore once the concepts of open innovation and open development become better integrated. Moving open development out of the constraints of the ICT4D domain will help researchers address key questions around open science, open education, and open data. ICT4D researchers should be proud of the open development theory that they have spawned. The next step is to let it loose.

Gender equity¹ is recognized by many to be critical to ensure an inclusive society that benefits all people. A gap in gender parity across the world has been well documented. The World Economic Forum’s The Global Gender Gap Report 2017 (Leopold, Ratcheva, and Zahidi 2017), which measures the gap between men and women in four key areas of health, education, economy, and politics, indicated that in 2017, an average gap of 32 percent remained to be closed between men and women. Some countries showed a reversal, and others some positive steps forward, indicating that context and sustained efforts are important while tracking gender measures. Yet these persistent gender gaps have focused global attention and stimulated a call for action to track progress and close these gaps (Leopold, Ratcheva, and Zahidi 2016).

Although the promise of open development is that it will be neutral, inclusive, and gender-fair, operationalizing openness to ensure that men and women benefit equally can be problematic. This is largely because gender asymmetries and the inherent social structures that fuel them are often not addressed adequately, even when technological and operational concerns are attended to. Addressing technological challenges faced by women is just that—a service delivery that reaches out to women. But participation is much more than numbers and attendance; it is about speaking up, sharing ideas, and being heard. Gender asymmetries in participation and voice have been attributed to a failure to address entrenched cultural and traditional biases and a lack of gender-sensitive policies (Neuman 2016). Participation is important for many open processes, but it does not happen on its own. Steps toward greater inclusion and participation involve gender-sensitive policymaking, strategies such as gender targeting (addressing the specific needs of women to ensure gender parity and representation), engagement with women’s civil society organizations, information flows that include women, and greater decision-making by women.

This chapter takes a closer look at inclusiveness and nondiscrimination in the open processes where, collaboratively, knowledge is produced, shared, and consumed. The argument is that knowledge itself is not gender neutral because it depends on who, what, and how it is created and used. To unpack inclusiveness, this chapter nuances the processes of participation and engagement in relation to gender—how participative, equitable, and gender responsive are open production, distribution, and consumption processes, and are they truly nondiscriminatory? To do so, it uses a gender transformative viewpoint that places issues of power at the center of the discussion. The argument is that it is important to recognize that all openness processes are affected by structural inequities and imbalances of power that engender social norms and interactions. Therefore, to be gender equitable, gender fair, and nondiscriminatory, one must address the power and structure that underpin open social processes.

The chapter progresses as follows. In the first section, it looks at open development processes through a gendered lens, examining the differences in men’s and women’s² experiences of open processes from the point of view of their socially constructed, ascriptive³ roles. This enables us to question the power relations between men and women which, in turn, intersects with power related to race, caste, class, abilities, and sexual orientation and keeps women in subordinate positions. Without using a gendered lens, one cannot address gender equity. In particular, it looks at the ways that open processes interact with and affect gender equity. Next, it describes gender concepts relevant to our discussion on open development and then explains how open processes (open production, open distribution, and open consumption of knowledge—see chapter 2 of this volume) can be engendered from or endowed with a gender perspective. The third section develops an analytical framework with guiding questions to determine the extent to which open processes are gender equitable, and then it illustrates the use of this framework with two case studies. The chapter concludes with recommendations for next steps.

Social inclusion is the removal of institutional barriers and the enhancement of incentives to increase the access of diverse individuals and groups to development opportunities (Dani 2003). Between the two extremes of inclusion and exclusion, there are shades and levels of both, which are largely dependent on institutional and social barriers. It is important to note that one may be included in one domain but excluded from another. In other words, social inclusion is a multidimensional, contextual and relational process capable of enhancing integration, cohesion, and solidarity (Silver 2015). Thus, the process of inclusion necessarily means that someone or something is or can be excluded.⁴

Historically, exclusion has been linked to gender, class, caste, age, religion, ethnicity, disability, and geolocation. While discussing his research on the use of digital technologies by Filipinos to improve transparency and accountability, Roberts (2017) points to digital technology emerging as a factor of exclusion. He identifies five A’s of digital access: availability, affordability, awareness, ability, and accessibility as a series of concentric circles that structurally exclude particular groups whenever digital technologies are deployed. Exclusion from and inclusion in social structures imply that someone has more power over others and that there is either an explicit or implicit power relationship influencing praxis (World Bank 2013).

The idea of inclusion is to some extent sameness—that is, everyone should be able to use the same facilities, take part in the same activities, and enjoy the same experiences, including people who have a disability or other disadvantage. The concept of sameness is recognized in formal gender equality, as it is premised on the principle of the sameness of women and men. It assumes that if women and men are given the same opportunity and women are treated similarly to men, equality will be achieved (Murthy and Kappen 2017). Thus, one way to ensure gendered inclusion (and nondiscrimination) is through this principle of sameness. Much of the research on differential access to the Internet or to open artefacts is related to this principle.

However, we could also consider a more substantive gender equality, which recognizes that men and women have to be treated differently to achieve equality because women’s disadvantages need to be addressed first, such as less mobility, lower buying power, different workloads such as child care, lower education, and lack of skills (Facio and Morgan 2009; Kabeer 2001; Kapur and Cossman 1993). Sen (1989) argues that outcomes cannot be judged in terms of access to resources, but rather people’s capabilities to engage in different functions (i.e., a disabled, or differently abled, person might require more support to engage in the same activity as an able-bodied person). But it is substantive gender equality that stresses that it is power and hierarchy (not just capabilities) that discriminate against women.

It is for these reasons that this chapter frames the discussion regarding inclusive open practices by addressing the socially and culturally constructed gender roles and relationships that often limit the capacity of women and men to participate on equal terms, even when open processes are available. We do this from a strong belief that open practices can be fundamental for societal transformation; but to better understand how this can happen, we need to understand the barriers to gender inclusiveness.

Analytical frameworks related to gender have attempted to assess to what extent and how we address gender equity. Displaying change on a continuum, the Gender Results Effectiveness Scale (GRES) analyzes the extent and process of structural transformation resulting from a project/intervention (UNDP 2015, 46). The scale prioritizes systemic change and suggests that access to opportunities and resources, or changes in laws and policies, is not enough to make a difference.

These categories necessarily evolve over time and are contextual. Gender-targeted or gender-responsive results have the potential to become transformative because they describe the differential needs of men and women and the unequal distribution of resources and rights. However, they are on a continuum and indicate the importance of moving forward toward the gender transformative category, which explicitly addresses the power structures and the roots of gender inequality.

Often, gendered responses have taken on a technocratic, apolitical form in an effort that appears to steer away from the structural challenges inherent in achieving gender equality.

Represented to technocrats and policymakers in the form of tools, frameworks, and mechanisms, “gender” appears as neutralised of political intent. Diluted, denatured, depoliticized, included everywhere as an afterthought, “gender” has become something everyone knows that they are supposed to do something about (Cornwall, Harrison, and Whitehead, 2004, 1).

Fundamental to any discussion on gender transformative change is the principle that patriarchal bias exists systemically and is manifest in structures (media, government, family, and other institutions), in interpersonal relationships, and daily lived experiences. In other words, gender permeates all that we do. As Hay (2012, 336) argues, “A gendered lens is recognition that the underlying structures and systems that create inequities cannot be programmed away within contexts that perpetrate and reinforce those systems.”

Applying a gender transformative approach to open processes would mean addressing gendered perspectives along with intersectionality, which refers to the intersection of gender with race, class, sexuality, ability, and other identities that marginalize and subordinate in the context of power. Such intersectionalities indicate how complex it is to address gender inequities. However, by using the gender transformative approach, open processes can demonstrate how the theoretical underpinnings of inclusion and nondiscrimination can be translated to practice.

A first step toward addressing gender and equity issues in openness is to better understand how access, participation, inclusion, and nondiscrimination are gendered, and how this affects decision-making and valuing knowledge that is created, shared, and consumed.

The Sustainable Development Goal (SDG) 5b on women’s empowerment has as one of its targets to “enhance the use of enabling technology, in particular information and communications technology [ICTs], to promote the empowerment of women” (UNCTAD, n.d., n.p.). Women’s empowerment is undoubtedly important for the attainment of this goal. Unlike the Millennium Development Goals (MDGs), which concluded in 2015, the SDGs are not just for the developing world; they are necessary for all nations, rich and poor, to address whatever inequities exist in every society. This means that we cannot talk about gender without talking about equity, and vice versa. Gender is only one axiom of inequity and cannot be abstracted from important identities, including race, ethnicity, class, culture, age, sexual identity, and physical ability. Gender (male, female, or other) injustice, a manifestation of exclusion and discrimination, is inextricably linked to these multiple identities. One may also argue whether it is possible to address equity without addressing gender, such as addressing poor people’s access to the Internet or the needs of Indigenous communities, or making Internet accessibility affordable. Such approaches focus on efficiency and effectiveness but are gender blind and assume that there is no difference in the position of men and women. But we know that women’s roles are gendered, ascribed by social roles, customs, and cultures that skew the power balance in favor of men. So, it is only when we align gender and equity that we are truly inclusive, we have ensured our commitment to having “no one left behind,” and we have addressed the broader, intersectional concerns of inclusion.

Past experience in the field of information and communication technologies for development (ICT4D) indicated that although there was a vision of being inclusive and gender-fair, in reality, men and women did not equally participate and benefit.⁵ At first, the slower inclusion of women was considered to be an artifact of digital access and paucity of skills. Then came the realization that the situation was more complex, and one needed to frame differences in terms of discrimination and barriers in participation that were structurally inherent in societies, often rooted in hierarchy and patriarchy. For example, Internet penetration rates among women from western Asia are lower than in any other place on the globe (Perryman and Arcos 2016). Citing research studies in Uganda and India, Perryman and Arcos (2016) show that women’s digital exclusion in some developing countries is due to oppressive gender-based norms, as well as low mobile phone ownership (in India, for instance, only 30 percent of mobile phones are owned by women). They observed that many of the inequalities posing a barrier to ICT use in the developing world replicate broader social inequalities: lack of education, poor ICT infrastructure, self-efficacy to use ICT, and lack of gender-sensitive policies (165–166).

Recent data on access and participation indicate a complex web of gender differences in the way that women participate in different countries and contexts.⁶ A recent set of household surveys found that, for the most part across the Global South, men have higher access to mobile phones than women, except in a few countries such as South Africa, Colombia, Paraguay, and Argentina. However, in all those countries except Argentina, men have a higher percentage of smartphones than women. Also interesting is that women in Ghana, Nigeria, South Africa, and Kenya use the Internet more for education than men do, and this effect is particularly strong for rural women in Nigeria. Similarly, in contrast to the rest of the countries surveyed, there are only two—Nigeria and Paraguay—where women use the Internet more than men for work-related activities (e.g., in Paraguay, 29 percent of women use the Internet for work, as opposed to only 16 percent of men). We need more data and analysis to understand why women do not possess high-value phones, and, when compared to men, why Internet and social media use is lower—what are the reasons for such differences, and how do they affect women?

Both men and women have reported online bullying and posting of inappropriate content, with more women reporting such complaints in Kenya and Tanzania and more men reporting them in Rwanda, Mozambique, Ghana, and South Africa.⁷ We need more research to understand the nature of the bullying and why it is gender specific.

Open development, typically predicated on ICTs, is subject to many of the same issues that emerged in the ICT4D field. In the same way that the diffusion of ICTs and their inclusion in development projects did not guarantee equitable benefits regardless of gender, open processes take place in the larger context of systemic inequities and do not necessarily transcend them. However, openness brings a particular subset of issues into focus and the normative impetus is gender equitable.

Looking through a gendered lens raises many questions regarding open processes and their underlying assumptions. For instance, in what ways are open practices engendered, and do those practices address hierarchical structures controlling knowledge production and dissemination? How and in what contexts are open development processes inclusive, nondiscriminatory, and therefore truly transformative in practice? When and how do open practices need to address power relationships that exist in society? What are the barriers for equitable participation of men and women across the intersections of class, ethnicity, status, and geographical location, and how do different openness processes address or affect them? Whose responsibility is it to ensure open and inclusive processes?

Open processes have distinguishing characteristics of being free, voluntary, and non-discriminatory (see chapter 2 for a more in-depth discussion of these characteristics). However, spotlighting these characteristics through a gender lens illuminates how open processes affect and are affected by the social differences between men and women and the power relations between them. Table 4.1 explains how engendering open processes can provide insights into the participation in and the development of knowledge artifacts.

Although open processes are based on technical principles of nondiscrimination, it is clear that they do not automatically mean that women will participate, or that women will participate in the same way as men (Buskens 2011). This is, of course, not surprising. What type of participation is valued (or not) is itself gendered—for example, good participation means that women participate with reference to their socially constructed, ascriptive roles or that they participate in ways that men do. But participation may not be good or valued if women participate, but do so differently from men.

To address gender and open processes meaningfully, we need a nuanced understanding of both access and participation. Participation cannot be understood as a binary function—either participation or nonparticipation. Rather, it is a continuum of various types of engagement and activity (Cousins and Whitmore 1998; Ramirez 2008).

A clear finding in the research is that access does not necessarily lead to participation. Access suggests the ability to use the information and the resources provided. Also, one may have access to resources but have no control over them, which means that participation is limited, particularly in decision-making. Graham (2014) and Warschauer (2002) acknowledge that greater access and participation driven by technological and operational innovations have not necessarily affected gender asymmetries, and they argue that much more work is needed to overcome inequalities in visibility, voice, and power in the networked society. A detailed discussion by Graham and De Sabbata (see chapter 5 of this volume) outlines the geographic asymmetries in openness participation. Information societies around the world are not free from existing and gendered global frameworks of governance, ownership, and control over resource access.

The analyses by Cornwall (2002) and Gaventa (2002) of how people participate are valuable for understanding how to frame participation in the digital world and, by extension, the social processes that define openness. Spaces for participation are important so that people can be directly involved and be able to exercise autonomous action. However, such spaces are produced spaces, never neutral, and reflect the interplay of existing power and difference (Gaventa 2002). It is no surprise that these spaces need to be situated not only within that context and its practices, but also with the multiplicity of other spaces with which they are in turn connected. Online participatory processes may serve simply to reproduce “echoes of dominant knowledge rather than to amplify the alternative, ‘bottom-up’ perspectives that are claimed for them” (Cornwall 2002, 9). All spaces have boundaries within which processes operate, acknowledging that “power must be analyzed as something that circulates,” meaning that we need to ask questions such as who is inviting participation and who is taking part, what they think participation is about, and how people in different spaces perceive it (Cornwall 2002; Foucault 1980, 102). A valuable framing of participation and power (Gaventa 2002) is expressed through the Power Cube and relates to how important it is to interrogate not only visible power (laws, rules) and hidden power (such as with exclusionary rules), but also invisible power (internalized self-limiting self-beliefs).⁸ Women and marginalized groups may not see their participation as valuable because they are told that it is not valued as much as that of men.

The concept of spaces of participation (and spaces of power) has value in our discussion of openness because no matter how open social processes are, they are always imbued with underlying status, class, and social position issues, which are reproduced in the ways that people communicate with each other in any social space, even those that are meant to be free (Kohn 2008). These prescriptors defining participation resonate with Lane’s discussion that educational divisions between people arise through a combination of factors—social, cultural, geographical, attitudinal, political, and economic (Lane 2009). These educational divisions can be easily extended to the digital world.

These issues are reflected in the data. For instance, a study of 3,705 responses of civic technology users who use mySociety’s open-source software extensively found the participation of women in civic applications such as FixMyStreet to be low. The fairly high gender imbalance among users diminishes women’s voices and marginalizes issues that are important to them –and consequently, the issues are less likely to be addressed (Rumbul 2015). Free and open-source software (FOSS) researchers Kuechler et al. (2012) make a comparison across a sample of open-source projects showing that although 25 percent of people employed in the information technology sector are women, only 7 percent of people who post online in open-source forums (i.e., post at least once) are women, and merely 2.5 percent of people who post regularly (i.e., ten or more contributions) were women. Women may have an online presence but do not contribute or speak out if they think differently from the online moderators, indicating a gendered pattern of participation that suggests women are “not supposed to” push back, assert themselves, or be otherwise expressive. A 2013 survey found that of more than 2,000 open-source developers who indicated their gender, only just over 10 percent were women (an increase from 2 to 5 percent in the early naughts) (Robles, Reina, Gonzalez-Barahona, and Dominguez, 2016). In this sense, the low and gendered participation is closely linked to autonomy and decision-making (World Wide Web Foundation 2015).

Studies have suggested that women tend to self-select what they do or do not do due to gender norms and social status rather than their technical skills and abilities. Some of these considerations, such as time use of women at work and home, have been well documented. Lam et al. (2011) confirm the presence of a large gender gap among editors and a corresponding gender-oriented disparity in the content of Wikipedia’s articles. They also hint at “a culture that may be resistant to female participation” (Lam et al. 2011). A Wikipedia Foundation survey found that only 13 percent of contributors were women and 9 percent of editors were women (Moeller 2009⁹; Wikipedia n.d.¹⁰). The reasons analyzed by Gardner (2011) elaborate on why women felt discouraged while contributing. Eight out of nine reasons cited by women had nothing to do with their skills but rather were related to women’s gendered roles and male privilege and dominance. The major reasons were not having enough free time, aversion to conflict and lengthy edit wars, belief that their contributions would be reverted or deleted, a misogynistic culture, sexual overtones that distress women, lack of comfort with using the male grammatical gender as part of language, and lack of self-confidence. This was followed by another study, Bear and Collier (2016), which explores the issue by analyzing a subset of the original 2008 Wikipedia survey data (1,589 occasional American contributors, of whom 17.5 percent were female) and finding clear differences along gender lines. Women reported feeling less confident about their expertise, being less comfortable with editing others’ work (a process that often involves conflict), and reacting more negatively to critical feedback than men (Bear and Collier 2016). There is already some research on the difference between men and women regarding confidence—but women do not actually score lower on ability and expertise (Torres 2016). Other studies have shown that while women are less confident than they should be, men are overconfident (Niederle and Vesterlund 2007).

Terrell et al. (2017) quote a study by Stack Overflow, a question-and-answer community for programmers, which found “a relatively ‘unhealthy’ community where women disengage sooner, although their activity levels are comparable to men’s.” A large-scale study of gender bias at GitHub shows that women’s acceptance rates of contributions are higher only when they are not identifiable as women. Their study indicates how deeply gendered responses are: women demonstrate a self-selection bias in how long they survive on the platform, but what was surprising is that there are different expectations for women’s and men’s work—women’s work is likely to be less valued or judged against higher standards than men’s.

What these data points illustrate is that, in practice, if we want openness to enable people to engage equitably, open spaces will have to address the particular needs of excluded groups and provide mediation between the various actors in that particular context.

Lessig (2003) questions whether the neutrality of the Internet is able to break boundaries influenced by power and ensure inclusion. But the reality is that knowledge institutions, producers, and distributors remain concentrated in a few locations, adding to inequity. The reasons Lessig identifies are the economic power of large knowledge institutions and the existing structure of intellectual property (IP) rights. Many organizations believed that correcting this imbalance was possible through a series of new-generation Internet tools that enabled global access to knowledge, as well as affording the opportunity to establish new institutions for sharing, peer production, and remixing of knowledge (Graham et al. 2011).

For instance, research on microwork has highlighted the complexity of addressing gender. Some argue that microwork and online jobs can benefit women because working online allows women to juggle family, work, and time use. A World Bank report (Rossotto, Kuek, and Paradi-Guilford, 2012) shows how microwork is successful in reshaping the global market, but it also points to a gender and age imbalance.¹¹ A recent study of a Latin American microwork platform by Galperin et al. (2015), however, shows how nuanced the gender dynamics can be. For example, overall, they found that, all else being equal, women were more likely to be hired over men (a small, but statistically significant effect). Yet women engage in microwork on the platform in smaller numbers than men and tend to submit lower bids for jobs than men do. However, when it is clear that women are doing the hiring, women will bid as much as men and will be chosen more often. There appears to be also a self-driven gendered perception which is nuanced across cultural perceptions of the value of men’s and women’s work.

Power is central to how gender affects open processes. Thus, it is crucial to address power and hierarchy if one expects everyone, including women, to participate in the open processes of production, consumption, and dissemination in an inclusive and nondiscriminatory way. If we address power differentials, which are often embedded in cultural norms and hierarchical values that impede women’s full participation, we will necessarily have to frame progress on gender in transformative terms. We can do so by ensuring that we address both the practical and strategic needs of women such as connectivity access, skills acquisition, and understanding dynamics related to family responsibilities. Addressing the strategic needs of women targets equity issues, such as education, mobility, economic participation, empowerment, and norms, which have the potential to transform gender relations. In other words, questioning, researching, and analyzing gender disparities and gendered differences of openness have the potential to ensure true inclusion and prevent discrimination in open processes.

Looking at how knowledge is produced, shared, and used, it is imperative to understand the knowledge economy (i.e., who produces and reproduces, who has access, and how people are represented and excluded). It is also about discussing inequalities in traditional knowledge and information geographies before moving on to examine the Internet’s potential to have new and more inclusionary patterns. However, studies show the divisiveness that exists in the knowledge economy. Graham (2014), in a study on the knowledge economy and digital labor, concludes that rather than democratizing platforms of knowledge sharing, the Internet seems to be generating a gendered digital division of labor, in which the visibility, voice, and power of the Global North are reinforced rather than diminished. Which knowledge system we value is clearly gendered—the knowledge system with more power is privileged over another, less powerful one. This is similar to modern knowledge being valued more than traditional ways of knowing (see also chapter 13 in this volume), or productive work being valued more than reproductive or caring work. Ann Weiner (2016, 1) writes that code is not neutral, as it is a creation, and that “[s]oftware products would be more powerful, more accessible, and more democratic—Twitter, for example, would look a lot different today if it had been built by people for whom online harassment is a real-life concern.” She claims that algorithms and code embody values and have been a concern for many years, quoting Ellen Ullman, one of the first women to enter the coding community in the 1970s, “[t]he engineer’s assumptions and presumptions are in the code” (Weiner 2016). Recent research acknowledges that gender bias pervades the open-source community (Terrell 2017). In interviews with women working in open source, Nafus (2012, 669 and 672) revealed that “men monopolize code authorship and simultaneously de-legitimize the kinds of social ties necessary to build mechanisms for women’s inclusion,” meaning that values are gendered, politeness is favored less by men, and “sexist behavior is … as constant as it is extreme.”

Rather than neutral and value-free, knowledge is deeply connected to a time, place, and social context. Feminist theorists talk about the situatedness of knowledge to a context and the effects of power on its production and validation (Brisolora, Seigart, and SenGupta 2014; Haraway 1988). This means that those who are marginalized have access to knowledge and viewpoints that are unique, not mainstream and dominant. Sandra Harding (1991, 185) asserts that such an approach is more objective and fair than traditional approaches to enquiry, which favor the dominant to shape knowledge, giving rise to a partial (and therefore distorted) view of reality. “Starting off research from women’s lives will generate less partial and distorted accounts not only of women’s lives but also of men’s lives and of the whole social order” (Harding 1993, 56). For knowledge to be constructed fairly, one must be aware of the skewed sociopolitical power and, more important, recognize that there are voices that have been left out and need to be heard.

This concept is important for openness because the cocreation (and cosharing and coconsuming) of knowledge assume that everyone has partial and contingent knowledge, and by participating in the crowdsourcing and peer-production processes, one achieves a shared knowledge. Similarly, perspective and experience shared through blogs and communities of practice provide important spaces for sharing, all adding to knowledge building. Knowledge sharing has become democratic (i.e., experts are not the only source of knowledge), valuing one’s experiences and thoughts, which can be created, accessed, and shared immediately. Here, one must recognize that it is filtered through the knower’s own realities, identities, and experiences, which shape the construction of shared knowledge. So knowledge has a social milieu and culture in which it is created and shared.

Situating knowledge helps us to be vigilant about who is not contributing to knowledge development. Masculine hierarchy can block and disregard the knowledge held by women. It is likely that the ones who are silent and have the least privilege may actually have the insight and knowledge that are valuable. Such knowledge may be traditional and gendered, and therefore less valued. Traditional bearers of local and Indigenous knowledge, be they men or women, may find themselves cut off from the networked society, where information, communication, and knowledge are tradeable goods. This is similar to the way that IP rights have taken away Indigenous ownership over traditional foods, medicines, and overall biodiversity (Thas 2008, 12). It has become essential that legal instruments are framed to recognize and protect knowledge created, developed, and enhanced by communities of people. These instruments need to be developed with the full participation of all parties who hold such knowledge, including men and women, and should acknowledge that men and women have differential access to the structures that shape knowledge systems (Primo 2003, 52).

Which knowledge is relevant and how it is expressed, therefore, depend upon one’s standpoint. The dominant paradigm is influenced by the politics of location combined with a specific scientific methodology (Baghramian and Carter 2017). Many researchers have commented that the dominant logic of the scientific knowledge paradigm (methods, writing style, and assumptions) is shaped by discourse that privileges dominant groups, such as the male European paradigm (Euro-androcentric), and shuts off alternative perspectives, typically held by less privileged groups. So, observations, expressions of lived realities, and the diversity of women’s experiences are often dismissed as subjective and irrelevant, and this overlooks the fact that there are often differences in how men and women think, what they think about, and what they consider important (Brisolara et al. 2014, 7). A project in the Open and Collaborative Science in Development Network (OCSDNet)¹² (called Understanding Open Hardware and Citizen Science) used language that was not part of mainstream science, such as the use of the word design rather than engineering. The project’s advisors from mainstream academia faced the dilemma of whether to use words that are commonly seen as “objective” (such as those commonly used in mainstream science) or to privilege women’s experiences (see chapter 13 in this volume for more on OCSDNet).¹³

So how does one navigate gendered open processes to create opportunities for systemic inclusion of marginalized or minority groups? In the next section, a framework is introduced that helps to illuminate gendered differences, which become apparent when a research framework exposes gendered patterns and contributing factors for those dynamics.

A gender analysis of open practices describes a systematic approach to examining factors related to gender and identifying and understanding the various roles, relationships, situations, benefits, constraints, needs, and interests of men and women in a given sociocultural context. Further research and analysis can contribute to understanding the nature of structural inequities in specific contexts and how they affect open practices. It is important to question preexisting differences that skew the distribution of positive impacts to men more than women, as well as analyze if the impacts themselves reinforce gender differences.

A number of researchers have explored the framing of gender analysis in relation to power relationships among men and women. Ineke Buskens looks at researcher intent, which she categorizes into three types: conformist (aims to produce knowledge), reformist (to produce knowledge in order to understand how to reform existing unequal gender relationships), and transformist (to enable women to produce knowledge themselves) (Roberts 2015). The latter type enables a better understanding for women about existing unequal gender relations and the structural power interests that support them so that these can be transformed.

In this section, I develop the novel Gender Analysis Framework for Openness (GAFO), which incorporates key questions that can help to assess the extent of gender inclusion in open development projects.¹⁴ The GAFO framework suggests that power is pervasive and can be described as power to, with, over, and within (Rowlands 1997), as well as power to empower (Chambers 2012). These categories can be described as follows:

•Power to: The increase in skills and capabilities so that one can contribute, decide, and take the lead. An example of how knowledge production can increase the power that comes from an OCSDNet project in Kyrgyzstan,¹⁵ which notes an increased participation of girls in citizen science in testing their communities’ water quality.

•Power with: Seeking collaborative and collective action for the collective good and to create an enabling environment. An OCSDNet project in Lebanon notes that the group of “community volunteers” testing well water were all women, and the team in Southeast Asia exploring open science through open hardware design workshops report that using the phrase “design and collaboration” drew more participation from women than did “tools and infrastructure” (see chapter 13 of this volume for more details).

•Power within: Leads to increase in motivation, confidence to contribute, sense of belief to bring change, excel, and lead change. Power within is usually expressed once power to, power over, and power with has been experienced. Both the previous examples, in Kyrgyzstan and Lebanon, indicate the power to confidently contribute.

•Power over: A person’s ability to overcome resource and power constraints to reach one’s potential and take control of one’s own personal and professional decisions and, in doing so, enable the person to increasingly influence and have a voice. The notion of cognitive justice used by the OCSDNet team from Haiti and Francophone Africa is one such example, which talks about empowered and confident researchers using all kinds of epistemologies and methods, not only those from the North (see chapter 13).

•Power to empower: Inspiring others, working on, and influencing broader agendas to multiply opportunities (i.e., being a champion). Examples can be found in the attempts of various forums to increase women’s representation and participation. The International Open Data Conference (IODC), a place for sharing experiences, networking, and discussing the most crucial issues of the movement toward open data, reported that only 34 percent of the speakers were women at its 2015 conference. This prompted an open letter to IODC 2016 from Mor Rubinstein, who called for gender-balanced panels for IODC 2018. “Let’s aspire (and commit) to have 50 percent women speakers this round” (Rubinstein 2016). It is not only important to see the increase in numbers, but also to review how men and women participate and engage in the open spaces provided for producing, reviewing, and using knowledge. Power to empower is possible only when there is power within to contribute. For example, Open Heroines created an online blog, providing a visible platform and public voice for women to express their thoughts, ideas, and critiques. This space emerged mainly because, despite women contributing to the open spaces, many felt their voices remained unheard, and were frustrated by this “underrepresentation.”¹⁶

In this section, I test the usefulness of the GAFO by using vignettes from various research projects supported by the International Development and Research Centre (IDRC)¹⁷ to understand how gender affects open processes. I chose relevant questions from the framework to describe gender dynamics, determine how gendered the open processes were, explain how power relationships were affected, analyze the various types of empowerment, and discuss the impact of doing so.

Case study: Teacher professional learning communities—A participatory open educational resource (OER) creation and adaptation approach in Karnataka state, India (a subproject of the ROER4D program)¹⁸

The Research on Open Educational Resources for Development (ROER4D)¹⁹ network aims to provide evidence-based research from a number of countries in the Global South (see chapter 12 of this volume). One ROER4D project in Karnataka used a bottom-up approach for teacher professional development, where teachers collaboratively and actively cocreated educational resources to respond to local needs.

The project worked with sixty-seven mathematics, science, and social science high school teachers and teacher educators in Karnataka. This group was embedded within a larger professional learning community of around 15,000 teachers across Karnataka, developed through the Subject Teacher Forum, an in-service teacher education program in the public education system. The research approach included periodic workshops with the sixty-seven participants, where they attended collaborative OER adoption processes. Analyses of data regarding the outcomes of the project indicated that teachers found meaning in reuse, creation, revision, remixing, and redistribution of resources on the mailing lists and the OER portal they used; experienced professional development and agency in using the digital platform; and increased their skills.

Questions from the GAFO (table 4.2) provide insight into the interrelationship between gender and openness. The project had to make special efforts to ensure that 50 percent of the teachers participating were women. Although female teachers are 41.42 percent of the total teacher population in Karnataka, only 20.9 percent (or a little over one in five) of the participant teachers in the OER project were women.²⁰ The selection process of science and math teacher resource persons was open and not discriminatory, and yet women’s participation was low. Here, it would be useful to enquire why many women lacked the power to. For example, the researchers might ask: What challenges do women face in order to participate and benefit from these open processes?

This OER project team made attempts to encourage women teachers to join in (power to empower) and to address the challenges faced by women teachers, both online and offline. We find that the participating women teachers had limited power with. While some women did participate in communicating through emails and sharing information from home, they contributed only 7 percent of all the emails; 93 percent of the emails were from male teachers.

The project explored a key power with question: What barriers do women face in communication and engagement with others in open processes – and/or in participating and/or leading collective or collaborative action? Women participants cited their various household and childcare responsibilities and a lack of time, preventing them from using their computers at home. Interestingly, male teachers reported that using the computers at home was more acceptable, and they did not categorize doing so as work. Such barriers affect the creation, sharing, and use of knowledge; limit participation in social processes; and, by extension, make them exactly the opposite of what they were meant to be—discriminatory and not inclusive. Although women who participated did show increasing confidence to participate (power within), the observation and gendered perspective of the project lead revealed greater nuance: women did not like conflict during discussions, were worried about being wrong, and were willing to go along with their male colleagues. The analysis indicated how gendered women’s (and men’s) responses to participation are, and, to ensure that open processes are transformative, strategies for inclusion, participation, and power have to be formulated right from the beginning.

The offline situations for the same group of participating women teachers included logistical barriers, such as the residential nature of the training program (which led to difficulties with childcare, seeking permission to travel from in-laws, and other issues); poor arrangements at the training venue, like dysfunctional toilets, lack of drinking water, and nonavailability of food arrangements; and public transport facilities close to the training venue. Some of the district training centers were far from the city center and not easily accessible. Women also mentioned that they did not feel safe at these centers. Those who did come said, “We came in spite of the home, we braved it” (power within and power over), indicating an attempt to exert control over barriers.

If we review this OER project on the gender effectiveness scale (see figure 4.1), the project could be classified as gender targeting—trying to get more women involved in the project by specifically addressing their practical gender needs. To be gender transformative, the project would need to question and address the power relations between men and women teachers and the roles that they played in their daily lives, which in turn affected their contribution to the OER development, sharing, and use.

Case study: Indigenous Knowledge and Climate Change Adaptation (a subproject of OCSDNet)²¹

OCSDNet explores the role of openness and collaboration in science as a transformative tool for development thinking and practice with researcher-practitioners from the Global South. The Indigenous Knowledge and Climate Change Adaptation project used a political and ecological approach to understand the relationship between climate change, IP, and Indigenous peoples, focusing on these facets as they pertain to Khoe peoples, and the Griqua and Nama groupings in particular, in South Africa. The project used participatory action research design and methods to reduce the power relations within and between researchers and the researched and hierarchies of knowledge production by involving marginalized groups within the design, implementation, and outcomes of the research. By doing so, these community researchers would be able to influence and contribute to the coproduction of knowledge.

Analyzing the case using the questions provided in the power to do category in table 4.2 helps reveal the role of gender in knowledge sharing and use in this project. Because men were the herders, they were de facto holders (or owners) of the knowledge about changes in the environment, and, within this group, it was the older and more experienced ones who held the power to do. In other words, gender roles influenced who possessed the knowledge; women did not have the power over to overcome the access to this knowledge, and neither could they have the power to nor the power within to access knowledge. However, there were some caveats to this scenario. For instance, some women had taken over herding, as they were widows or because their husbands were migrating for work. In these cases, they did possess knowledge, and had the power to contribute. But researchers noted that women were not comfortable coming to meetings, and, even if they did come, the male elders preferred to talk to researchers and express their views to them. Women had little power within or power over to overcome the barriers. In fact, it was easier to have more interactive conversations separately with women.²² Clearly, the filtration of Indigenous knowledge, despite its coproduction, is still gendered. Researchers involved in this project mentioned that to explore the role of gender deeply, it is important to have more time, resources, and expertise at their disposal.

If we review the project using the Gender Results Effectiveness Scale (see figure 4.1), it could be considered gender discriminatory because it did not question at all the place of women in sharing the knowledge. In fact, the researchers in the example noted that men were given precedence, even though there were circumstances where single women herders were the holders of knowledge. To be gender transformative, they had to challenge the male hierarchy so that women could participate in joint meetings, have their voices heard and valued, and reconstruct Indigenous knowledge through a gender lens. This is possible only if the research design and openness values of inclusiveness specifically articulate such questions.

As can be seen from the examples given here, many criteria need to be considered in making openness inclusive and nondiscriminatory. Open spaces are not neutral, and definitely not gender neutral. In some cases, these spaces are pointedly gender discriminatory. The following sections attempt to address how open spaces can be gender-fair and safe.

A few of the ROER4D projects have shown that women need much more than mere persuasion and open space to participate. For example, Maria Pilar Saenz Rodriguez from the Colombia OER project spoke during her interview about the need for a collaborative and supportive space, where women teachers can talk freely about their personal lives, as well as work professionally.²³

Some other OER project leaders also noted a similar trend during interviews; Lauryn Oates and Mubarak from Afghanistan²⁴ mentioned that in their OER project, men and women had separate times to access a computer lab, which is itself telling. Men tended to dominate discussions, even though teachers were university educated and the training was located in capital cities, where a more equitable attitude was expected. There were more male teachers in Afghanistan, but the project had ensured a 50:50 representation. Similar to the Karnataka project, there was a difference in the online participation of men and women, with more men uploading the material and using the data. Men were more used to the technology and stayed longer to use the computers, unlike the women.

But gender can work the other way, too, as happened in an OER project in Sri Lanka, where female teachers in Colombo managed to attend more workshops while juggling childcare, whereas male teachers preferred to pursue other degrees where there was greater monetary gain.²⁵ This example indicates that participation is often nuanced and contextual, and the perception of benefit can differ between men and women.

Generally, as power structures get disrupted due to increases in inclusion and participation of women in development, we are warned of the possibility of a backlash. This is equally true of digital spaces.

Susan Herring’s research on gender dynamics in participation in online discussion lists provides plausible reasons for the gender gap reflected in the 2011 Wikipedia editor’s survey; more women than men gave their main reason for not participating in online posting as that they felt unsettled by the “tone of the discussions” and “antagonistic exchanges” (Herring 2011).²⁶ Similarly, Reagle (2011) points to the presence of a “culture of hacker elitism,” which can deter female contributors from utilizing free culture projects like Wikipedia and free and open-source software. Reagle (2011) suggests that the “ideology and rhetoric of freedom and openness can then be used (1) to suppress concerns about inappropriate or offensive speech as ‘censorship’ and (2) to rationalize low female participation as simply a matter of their personal preference and choice.”²⁷

Linked to safe participation is the need to promote digital rights, which is not a very commonly known set of human rights issues, especially in the South Asian context. At the Hamara Internet—Ending Online Violence against Women conference, held in Pakistan during November 2016, Shumaila Jaffrey of the British Broadcasting Corporation (BBC) referred to a recent report of the Federal Investigation Agency on cybercrimes, which suggested that 45 percent of the victims in the 3,000 cases of online harassment in Pakistan were females.²⁸ Thus, digital security and protection is critically important in our discussions of gendered inclusion in open processes. Buskens’s (2011, 72–73) experience with the Gender Research in Africa and Arab Countries into ICTs for Empowerment (GRACE) project indicates that power dynamics in the environment have the potential to corrupt “the quality of openness,” despite the intentions of stakeholders and role players.²⁹ She points out that “even within this network space, open sharing between the researchers had to be mediated by their need for safety. There was an awareness that. … GRACE researchers are also members of many social systems in their countries of origin, as well as regionally and internationally, which are very diverse, and not all share enlightened perspectives on women’s empowerment and gender equality. The GRACE social platforms were therefore managed through a rhythm of openings and closures” (Buskens 2011, 74).

There is a need to address the context and the sociocultural milieu so that women are motivated to contribute to the cocreation of knowledge without being threatened. For example, on realizing the low number of female editors, Wikipedia initiated “edit-a-thons,” with the idea of (1) increasing the coverage of topics related to women in Wikipedia; (2) encouraging more women to edit Wikipedia through projects like VisualEditor; and (3) providing a user-friendly environment for female newcomers through the Teahouse project. The Wikipedia initiatives show the possibility of having two strategies: (1) women need to be targeted separately from men, and (2) the setting should include both men and women and be sensitive to women’s concerns. Both strategies have their advantages and disadvantages and can be used creatively by other organizations besides Wikipedia.³⁰

These examples exemplify power to empower to provide safe online space to women contributors. Women’s power to contribute, to be able to exercise power over online harassment, and the power within to challenge safety issues link to the work-in-progress GAFO framework.

Whether a radical idea such as the feminist Internet (Association of Progressive Communication n.d.)³¹ is the answer, or if there must be rules that protect women in social praxis until cyberprotection is ensured, discussion of openness as being nondiscriminatory is limited.

Roberts (2017)³² discusses how, by focusing on the excluding mechanisms, designers of digital development projects can use different and imaginative ways to seek the participation of marginalized groups, including blending offline and online activities and using analog as well as digital technologies. He underscores that if there is an intent to include the most marginalized in digital development initiatives, then there is a “need to design for equity from the outset.”

While openness implies a normative principle of inclusion, if we do not explicitly populate it with inclusion principles, it will only reflect the dominant paradigms existing in society. What is evident is that if gender is not addressed specifically, any discussion on openness is likely to bypass it—that is, “If you don’t ask about gender, you don’t learn about gender” (ROER4D 2016). Most likely, the knowledge created, accessed, and shared will reflect the dominant experience, which is likely to be male and white. To reach truly equitable representation, it is clear that one has to create conditions that address the barriers that prevent participation.

At the simplest, one must assess the differential use rates of men and women. It is important to then assess the direct impact on women (in terms of communication, decisions, and information); compare the impact among men and women, in terms of the relationships and the gendered actions, roles, and resources available; and finally examine the forces (sociocultural, institutional, and organizational) that influence gender (and gendered) norms, power, and practices (Heeks and Molla 2009).

Since openness, which is about production, sharing, and use, is ultimately a collaborative endeavor, it is often necessary to accommodate dynamic teams that are more geographically distributed. This is important, as it is one of the more positive and significant predictors of productivity (Terrell et al. 2017). Bias mitigation activities can be useful here, such as (1) “bias busting” workshops,³³ (2) open-source codes of conduct,³⁴ (3) blinded interviewing,³⁵ and (4) acknowledgments by the community that biases are widespread, so that it can make a practical impact on the practices of open development (Terrell et al. 2017).

The need to create and nurture a safe and nonbullying atmosphere for the Internet in general, and open processes specifically, as a global concern cannot be overstated.

It is imperative to tackle barriers to women’s empowerment by enhancing their participation and representation in decision-making processes at all levels. Only this long-term process can have a sustained impact on developing a gendered Internet space and, therefore, on openness.

However, several steps and strategies may be initiated so that this extensive endeavor becomes manageable. The following actions are suggested for centering gender in openness. First, as one’s gendered orientation will determine whether gender will be addressed in open development, it is important to build the capacities of all those involved in open production, cocreation, sharing, and usage to review processes using a gender lens. If you put on a different lens, you can ask: “How is this issue for women?” “How is it different for men and women?” “Why and how can we change it?” Second, having gender consultants involved in the open development process of a project could be a short-term way out, but the long-term approach of training all those involved and creating a gender equal work ethos would be more sustainable. It is essential to improve the capacity of program leaders and partners to develop and scale up gender-responsive programming to advance gender-related outcomes in program areas. Third, the inclusion of team members and consultants, trained for gender analysis, needs to be accompanied by the appropriate resources for developing a strategy for engendering openness—right from the start. Fourth, there is a need to address women’s practical needs and expand safe spaces for voices, choices, and access; good practices like HarassMap and the Government of Canada’s commitment to feminist strategies in international development need more visibility. Fifth, we must encourage good practices and critical feedback from reviewers and evaluators to remove the defensiveness within networked economies about gender-related concerns.

Gender equity in openness needs much more than just wishful thinking. It requires a head-on discussion regarding the power interplay, explicit and implicit, that colors gender in open development.

Even though the Internet is theoretically open to anyone, it is characterized by deep geographic inequalities. Some people and places are far more likely to participate or be represented than others. This chapter explores the geography of participation and representation in open platforms across the globe. It begins by looking at the geography of Internet penetration as a prerequisite to participation, exploring which parts of the world are better connected than others. It then moves to an in-depth exploration of two types of participation in open platforms. First, it looks at the geographies of participation through Wikipedia (as a proxy for content generation) and GitHub (as a proxy for code development). Second, it looks at the geographies of representation (in other words, the parts of the world that have more or less data about them) through OpenStreetMap, the world’s largest open map, and GeoNames, the world’s most used open gazetteer. The discussion asks which Internet users are more likely to participate in the global web of knowledge than others. Once technological barriers to participation (i.e., Internet access) have been bridged, why do significant barriers remain for large parts of the planet?

Information has always had a geography. It is from somewhere, about somewhere; it evolves and is transformed somewhere; and it is mediated by networks, infrastructures, and technological tools, all of which exist in physical, material places. Information has also always been mobile. The most commonly used definition of information, as noted by the Oxford English Dictionary (2015), is “the imparting of knowledge.” It emphasizes the transmission and movement of information in an ongoing and dynamic process. In short, information, ranging from standardized measurements to instruction manuals to stories and legends, emerges from and is adapted to a range of local contexts and geographies.

If information has a geography, then so does open information. In the context of this research, we define open information (or open informational resources) as information that is freely and publicly shared, typically (but not exclusively) over the Internet. Some individuals, groups, networks, and institutions are more likely to produce and share open information than others. Also, some parts of the world are more likely to have more open information about them than others.

In theory, open information is equally accessible and usable by all, and therefore portends more equality with respect to the meaningful use of this information. However, if information itself is characterized by large inequalities in terms of both who produces and what and where it is produced, then this will likely translate into large inequalities in terms of who can make meaningful use of open information. The idea that better access to technology and to informational resources and open information might be a force for inequality as well as equality is not new to the literature (e.g., Zook et al. 2010; Foster and Graham 2016). This chapter specifically focuses on the geographies of open information as a component or function of inclusion in and exclusion from the meaningful use of open information.

In focusing on the information geography of open information, this chapter specifically accesses patterns of mutability (i.e., the tendency for the information to be changed), mobility, and underlying power relations relative to historical arrangements in the production and use of information. It is precisely this mobility and mutability of information that provide the motivation to constrain the mutability of information through the creation of what Latour (1986) refers to as “immutable mobiles” (i.e., information that can be transported without significant change to its inherent characteristics or meaning). The printing press, for instance, made it relatively cheap to create and transport information within the container of printed paper, while simultaneously limiting how its form could change. As information technology–based nonproximate communication emerged and was adopted by governments and companies, the ability to create immutable mobiles, or shared understandings of information by people separated in both time and space, became vital. Information and power thus became intimately intertwined as people capitalized on the value associated with epistemic control: information represented this and not that (Schech 2002; Foucault 2000).

A key characteristic of immutable mobiles is their ability to crystalize informational layers of places in a moveable container in order to create particular geographies of information (Wilson 2015). A map, a tourist guide, and a postcard all annotate a bounded part of the world tied to a relatively immutable form that can be physically moved to different locations. Thus, immutable mobiles affect representations of places that are used both in the locales referenced and in other parts of the world (Kitchin and Dodge 2007; Kitchin, Dodge, and Perkins 2011). Information about any location can be abstracted from an immutable mobile and be placed in the relatively immutable form of printed paper, thus simultaneously fixing information to a physical object and untethering it to a locale, as the printed form can be easily moved through space. In the era of print, the geography of immutable mobiles such as maps or books largely defined the geography of codified and geographically referenced information. In other words, the scope for the spread of open information was always constrained by the geographies of immutable mobiles that transported and contained it.

Information communication technologies (ICTs) have also facilitated an evolution of information beyond immutable mobiles through the creation of (im)mutable augmentations characterized by the layering of dynamic information across and over geographic space. Thus, not only does information have particular geographies, but geography itself is layered, defined, and augmented by information that is more or less immutable (i.e., both mutable and immutable, or (im)mutable) depending upon the institutions and practices associated with it (Perkins 2014). The Taj Mahal, for instance, is not just a building made from marble and mortar, nor is it simply represented by guidebook entries, postcards, or other immutable mobiles that are trapped within their containers. It is also overlaid with digital images, videos, descriptions, reviews about tours and past performances, and innumerable stories told about experiences associated with it, which are stored in online maps, annotations, and websites.

All those things are informational, but they are also part of the place itself; they are part of how we enact and bring the place into being (Graham 2013; Graham, Zook, and Boulton 2013; Floridi 2011).¹ The advent of the (im)mutable augmentations and mobile tools that allow us to access this information while in situ (for shopping, way-finding, driving, sightseeing, protesting, and many other geographic activities) places an ever-greater value in the epistemic control to fix informational layers of place (Shaw and Graham 2017). The fact that contemporary (im)mutable augmentations are digital means that they hold much potential with respect to open information. In theory, anyone with access to the right tools and modes of connectivity can produce, share, and access a world of information.

Geographies of the production of data matter because they can reinforce a world in which some people have a more substantial voice than others. In a world in which digital information is even more crucial to almost every facet of daily life, it is crucial to understand whether some groups of people have more of a say in how our digital world is constructed than others do. The question that could be asked is: What sort of global information society are we building if large groups of people rarely participate in it as producers? Furthermore, the (im)mutable augmentations of information about places matter because they shape how we are able to find and understand different parts of the world (Shelton et al. 2014). Places that are invisible or discounted in representations are equally invisible in practice to many people. A restaurant omitted from a map can cease to be a restaurant if no one finds it. Likewise, how places are presented within informational augmentations fundamentally affects how they are used or brought into being (Graham and Zook 2013). In other words, geographic augmentations are much more than just representations of places. They are part of the places themselves, and they shape it rather than simply reflect it; the map again becomes part of the territory (Floridi 2014; Pickles 2004). This fusing of the spatial and informational augmentations that are (im)mutable can result in high stakes as annotations of place emerge as sites of political contestation, with different groups of people trying to impose different narratives on informational augmentations (Zook and Graham 2007).

In short, the geography of information, which refers to the geographic distribution of information either as a phenomenon in its own right or as a representation of some other underlying process (Wilson 2015), has long been a key means of control and power formation. The rise of information geographies (i.e., informational augmentations to places) represents a key and emerging area of inquiry for scholars of information.

After giving a brief review of predigital geographies of information, this chapter explores how open information geographies have their own geographic distributions: first in terms of geographies of participation (contributing content to a digital platform), and second in terms of representation (what informational augmentations exist about a place). The chapter does this by examining platforms that play a particularly important role in the production and sharing of open information and information-embedded goods—and thus play a critical role in the ecosystem and political economy of information flows.

In the predigital age, the affordances of technologies and associated sociotechnical systems for collecting, storing, and disseminating information meant that information was both scarce and geographically embedded. For instance, at the dawn of the nineteenth century, the tools for collating encyclopedic information about places (e.g., compasses, paper, and sextants) were concentrated with just a few individuals in a few places, making the ability to engage in large-scale data collection relatively rare. This was even more the case because the required propinquity to the object of measurement inherent in collecting information about places entailed a scale of organization not widely available. Likewise, the state of the art for containing information—the book—required specific points of access to codified content for its production. Books were also constrained by a particular form—two-dimensional printed material of prespecified parameters, which is usually part of a linear reading trajectory from start to finish, and features a method of physical storage in particular places.

Other key bottlenecks in the processing of information included the availability of requisite skill sets to manage, validate, merge, modify, classify, sort, analyze, and manipulate information into particular forms and formats. The institutions that emerged to meet this challenge, such as universities, associations, and guilds, required resources. These in turn often needed to be spatially fixed as a result, and a formalized education in information-handling practices was necessarily place-bound and concentrated in specific locales. These geographic centers of calculation meant that the ability to access codified information, let alone contribute to it (i.e., the professionalization of knowledge work), was also highly constrained.

Although the affordances of predigital technologies (and their associated systems of governance, economization, and socialization) allowed the movement of codified information (shifted from its point of creation to other locations), it could never transcend the innate materiality of its medium or the world. In short, the frictions of mobility associated with transmitting and storing information, the place-bound rules and forms of governance, and the availability of requisite technologies have all shaped the geographies of information in this era. In practice, these constraints manifest into hegemonic representations and hegemonic modes of participation (cf. Gramsci 1971). As knowledge and codified information are always produced under conditions of power (Crampton 2008; Pickles 1994), control over hegemonic representations has been a way of exerting economic, social, and political power (Laclau and Mouffe 2001).

Books, newspapers, and patents, for instance, were all far more likely to be published from (and about) the Global North, with the Global South playing a relatively minor role in producing, using, and controlling codified information (Zhang et al. 2013; Graham et al. 2011; Thompson and Fox-Kean 2005).² This information power and power over information manifest in distinct spatial patterns and almost all traditional media of information are characterized by significant spatial inequalities, leading Castells (1999, 3) to conclude that “most of Africa is being left in a technological apartheid.” The systemic and uneven relationships between information production/use and socioeconomic exclusion and marginalization are characterized as the “black holes of informational capitalism” (Castells 1997, 162).

Against this backdrop, the past decades have seen a sea change in the availability of information. The terms information revolution (Floridi 2014) and data revolution (Kitchin 2014) signify the radical changes in the ways that information is produced and used, implying that we are now in an age of postinformation scarcity. In this age, information has moved from being proprietary and closed to becoming shared and shareable, and it no longer is confined to a narrow selection of containers and places. This transformation has been brought about by the proliferation of new sociotechnical systems of the so-called information age. These systems are underpinned by greater accessibility to computers that can readily receive and transmit information nonproximately, as well as by a host of associated social, economic, and political practices. As a result, many of the barriers to the production, processing, and proliferation of information in the predigital era have been drastically lowered.

Some have argued that this emerging digital age offers a potentially radically different political economy of information³ (Benkler 2007; Bruns 2008; Jenkins 2006). Tapscott and Williams (2006) and Shirky (2011) highlight the ways that digitally mediated participation and representation are broad-based, circumventing traditional mediators of information that allow citizens to play a more significant role in shaping the content and augmentations that play key roles in their lives (also see, e.g., Sui and Goodchild 2011). Lawrence Lessig, a key scholar of the effects of information systems on property rights, has made some particularly hopeful observations on the democratic power of the Internet. At the 2003 World Summit on the Information Society, he remarked on the significant possibilities afforded by the web, noting that “[f]or the first time in a millennium, we have a technology to equalize the opportunity that people have to access and participate in the construction of knowledge and culture, regardless of their geographic placing” (Lessig 2003, 1).

Lessig’s characterization is not unique; such sentiments also infuse the worlds of policy and business. In 2012, Hamadoun Touré, the secretary-general of the International Telecommunication Union (ITU), claimed that once Internet connectivity arrives, “all the world’s citizens will have the potential to access unlimited knowledge, to express themselves freely, and to contribute to and enjoy the benefits of the knowledge society” (Touré 2012). Echoing similar lines of rhetoric, Wikipedia, according to its cofounder Jimmy Wales, seeks to “contain the sum of all human knowledge” (Miller 2004). And Google’s stated core mission is to “organize the world’s information and make it universally accessible and useful” (Google 2015).

Similar rhetoric has accompanied the emergence of openness whose influence has piggybacked on the growth and diffusion of ICTs. There were hopes that openness would make a difference to practices and processes of development at the world’s economic margins (Bentley and Chib 2016; Graham and Haarstad 2011; Smith 2013). This is in part because the global or macro picture is one of extreme connectivity—currently, there are 3.5 billion people and 25 billion devices attached to the global network. Furthermore, the total amount of accessible information online, both open and closed, has dramatically increased (Hilbert 2012). But has this increase in access and availability of freely available open information had any effect on the political economy or geographies of information? Has it altered participation or representation? If people can, in theory, create and access open information about and from almost anywhere on Earth, are we seeing different geographies of participation and representation or new layers of digital augmentations [i.e., data shadows (Graham 2010)] associated with places?

To answer these questions, we engaged in a broad-scale survey of the contemporary geographies of key digital and Internet-mediated platforms for open information. Some of this information is produced through the open production practices of crowdsourcing or peer production. We produced a series of mappings based on the most definitive sources available. Some data is emerging from established informants such as the ITU, which provides relevant indicators of information use (and as a result, these data have been fairly widely disseminated). By contrast, much of the data used in this chapter were generated using bespoke data collection tools making for a unique analysis of rarely used data.

To understand the geographies of participation, we explore GitHub (as a measure of contributions to the world’s largest open software repository) and edits to Wikipedia (as a measure of contributions to the world’s largest open encyclopedia). While both measures are shaped in part by local cultural practices, such as the awareness of and interest in contributing to GitHub or Wikipedia, they also represent global communities of practice that transcend the particularities of any one place.

GitHub is the largest code-hosting service in the world, containing 17 million repositories. In 2015, it had 3.5 million registered users, as well as countless unregistered users. Programmers can use GitHub to publish their code for others to download and use, as well as to collaborate with others on shared projects, tracking changes and contributions. No other code-hosting service has anywhere near the same number of users, making GitHub a useful proxy for the extent to which Internet users create and share code.

Data for the map of GitHub users and commits (i.e., contributions of software or code for others to use) in figure 5.1 come from the GitHub archive (GitHub 2014) and include all freely shared contributions logged by the service. The 2013 data contain over 65 million commits made by about 1.1 million users who were active that year. Users are able to specify their home locations as a string of text, and 26 percent of users’ entries were geolocatable using the Edinburgh Geoparser (2015) web service at the University of Edinburgh. Another 2 percent of entries were given in nongeographic terms (e.g., “Internet,” “127.0.0.1,” “Planet Earth,” or “everywhere”), with the remaining 72 percent of profile entries being empty. The 26 percent of locatable users account for over 44 percent of the commits. While this is a relatively small percentage of all users, we are not aware of any systemic bias that would significantly overcount or undercount contributions from particular countries.

Figure 5.1 shows the number of GitHub users and commits by these users relative to the number of Internet users. While Internet penetration explains about one-third of the variability in the number of GitHub users per country, large differences remain in the rate at which people from various parts of the world contribute. North America and Europe each account for over one-third of the total number of GitHub users, and the rate of participation—34 GitHub users and 21 GitHub users per 100,000 Internet users (hereafter referenced as G per 100 K), respectively—well exceeds the global average of 11.4 G per 100 K. The platform is particularly popular in northern Europe and eastern Europe; for instance, Iceland and Sweden have more than 50 G per 100 K apiece. It is also popular in New Zealand and Australia, with about 35 G per 100 K.

A majority of the remaining third of GitHub users (17 percent) are located in Asia, including Singapore, with 27 G per 100 K, and Taiwan, with 10 G per 100 K, which per capita represent the two largest. In absolute terms, China is home to 5.6 percent of GitHub users but has fewer than 3 Github users per 100 K. The Middle East, North Africa, and sub-Saharan Africa stand out with the lowest levels of participation, and, combined, the regions are home to less than 1 percent of GitHub users and commits, comparable to the level of activity in Switzerland.

In addition to being home to the majority of users, North America and Europe contribute at higher rates than the rest of the world. For example, the United States is home to 31 percent of users but contributes 35 percent of commits. The statistics for users and commits for other standout countries include the Netherlands (1.7/2.4) and Switzerland (0.9/1.4). The opposite pattern is evident in much the rest of the world, with India accounting for 3.6 percent of users but only 1.7 percent of commits; Brazil, with 3.7 percent of users and only 2.5 percent of commits; and South Africa, containing 0.46 percent of users but only 0.36 percent of all commits.

In summary, this analysis of GitHub shows that Europe and North America are significantly more involved in collaborative code development than other parts of the world. Africa and the Middle East contain considerably fewer people accessing or contributing software than would be expected, given the level of Internet access in those places. One potential inference from these data is that factors such as skills, education, and income play a more significant role than connectivity (Ojanperä et al. 2017).

While code is central to the functioning of contemporary societies, its creation is a particularly specialized endeavor. In contrast, Wikipedia—a platform that allows all Internet users to write and edit its articles—offers an indicator of a more generalized level of content creation and contribution. Not only is Wikipedia by far the world’s largest and most used encyclopedia (1,600 times larger than the Encyclopedia Britannica), it is also extremely popular. More than 15 percent of Internet users access it on any given day, and it is in the top twenty most accessed websites in 95 percent of the world’s countries (Graham et al. 2014). Part of its popularity can be tied to its articles being written in 282 languages, with 40 of those language versions containing more than 100,000 articles each. In other words, edits to Wikipedia have a tremendous power to shape the content accessed by the majority of Internet users around the world.

Wikipedia’s approach to contributions, in theory, allows anyone with an Internet connection to contribute; the platform’s strapline is “the free encyclopedia that anyone can edit.” However, based on the number of edits to every language version of Wikipedia coming from all countries and territories in the last quarter of 2014 (the most recent data available),⁴ figure 5.2 shows that the geography of participation on Wikipedia is highly uneven.

Stark inequalities are readily apparent, as Europe and North America contribute 35.2 percent and 23.6 percent of Wikipedia’s edits, respectively, while Africa as a whole contributes only 1.3 percent of the world’s total. In fact, contributions from Africa are so low that there are actually more edits that originate in the Netherlands than the whole continent. While some of these disparities can be explained by the total number of Internet users in a country, even normalizing by the percentage of the population online, far fewer edits are registered in Africa than would be expected (for a detailed statistical analysis on the topic, see Graham et al. 2014).

A key way that these geographies of participation matter is tied to how content and locales or places are created in different parts of the world. Figure 5.3 shows the percentage of local edits to articles that reference places within the users’ own countries.⁵ In other words, this map illustrates the percentage of edits about a country (i.e., articles that fall within the boundary of a country) that come from users within that country. In this specific instance, we were only able to use data about the English version of Wikipedia because the method of geolocating user profiles involved parsing unstructured English-language text. While many places have a high percentage of content created by local people (e.g., 85 percent and 78 percent of content about the United States and the United Kingdom, respectively, comes from users located there), the percentages in Africa are much lower. Only 16 percent of content about Nigeria and 9 percent of content about Kenya are created by locals, and, in much of the rest of the continent, less than 5 percent of content is generated locally.

We can also analyze the locality of voice and participation using figure 5.4 (originally published in Graham et al. 2016). It depicts seven world regions (Europe, Oceania, North America, Middle East and North Africa, Latin America and the Caribbean, sub-Saharan Africa, and Asia) as dots. The y-axis of the graph indicates the proportion of edits that a region receives internally (a high score would mean that the region largely is written about from within the region). The x-axis indicates the proportion of edits that a region commits internally (a high score means that most edits from a region are about that region). It is worth noting that even though a region like sub-Saharan Africa and a region like Europe commit roughly the same proportion of edits internally, sub-Saharan Africa has far fewer locally produced digital representations than Europe. This is, in part, because, as presented in figure 5.4, there are far more editors from Europe, and it only takes a few editors writing about sub-Saharan Africa to result in articles about the region being dominated by nonlocally produced content. Compounding this problem is the fact that editors in other parts of the world with very limited locally produced content (such as in the Middle East) end up exporting a lot of their editing power by writing about other parts of the world.

In summary, this section shows the spatial distribution of participation in two open production processes. While both data sources have their own idiosyncrasies and display different patterns (to be explored in future research), the general refrain remains consistent. Thus, the Global North is characterized by the greatest levels of participation and is creating the bulk of digital content, while the Global South contributes very little. The continent of Africa, in particular, is almost entirely omitted from these processes of digital generativity. Moreover, these distributions stand in marked contrast to the geographies of Internet users. While this is cause for concern in a number of arenas, we now turn to a direct impact of these differences—namely, the ways in which places are represented (or not) within information geographies.

To better understand the contours of representation within the (im)mutable augmentations of the world, we present a series of visualizations (i.e., we ask what informational augmentations exist about a place): OpenStreetMap (OSM) (the world’s largest open map) and the Semantic Web, as defined by GeoNames (one of the world’s largest open gazetteers), both of which are examples of the crowdsourcing of information. In both of these cases, the popularity of these systems provides useful ways to gauge the spatial differences in digital representation on these crowdsourcing platforms.

As the world’s largest road map, and because it is free and open source, OSM is used as a base map for thousands of other digital platforms and services. Therefore, it is important to understand how much content OSM contains about various parts of the world.⁶ Figure 5.5 is based on data downloaded from GeoFabrik.de in December 2013, containing all 2 billion nodes (i.e., elements used to represent any point feature) in OSM at that time. The number of nodes in each cell of the 0.1-degree grid were tallied and mapped. The node density was then calculated for square area units with side length of 10 kilometers and a neighborhood radius of 25 kilometers. The United States accounts for the largest total amount of content, hosting 21 percent of all nodes present in OSM, followed by France, Canada, Germany, and Russia, with each containing more than 100 million nodes. These five countries alone comprise 58 percent of all the content in OSM, and the group of 36 high-income Organisation for Economic Co-operation and Development countries contains almost 80 percent of the total number of nodes.

The Netherlands has the highest density of content, with an average of over 1,000 nodes per square kilometer. Belgium follows with over 700 nodes per square kilometer. Germany, the Czech Republic, Switzerland, and France all have about 400 nodes per square kilometer. In contrast to the high density in Europe, the Southern Hemisphere is far more sparsely covered. Africa and Latin America are each represented by less than 5 percent of the world’s content in OSM. In contrast, California alone accounts for almost as much content as the whole continent of Africa. Furthermore, Iceland has as much content as Egypt, despite being a tenth of its size and having 0.4 percent of its population.

The Semantic Web is a movement to make the web more efficient by creating common frameworks that allow data to be easily shared and reused across sites, services, and places. Google, for instance, has recently employed what it terms its knowledge graph as a way of reacting to user queries. Instead of displaying a simple list of links to websites in response to a search query, the search engine looks for implicit meaning embedded in a request (such as “What is the capital of France?”) and then displays structured information about that query on the right side of the search page (e.g., the word Paris, the population of Paris). Because Google is the primary interface to the web for approximately 67.6 percent of Americans and over 90 percent of Europeans (European Commission 2013), information in its knowledge graph plays an important role in how hundreds of millions of people view and interpret the world.

Information contained in the Semantic Web (such as the answer that Paris is the capital of France) is derived from centralized databases that feed facts into Internet services like Google, Bing, Facebook, and many others. One of the most important back-end geographic databases for the Semantic Web is GeoNames, the world’s largest and probably most used gazetteer or directory of place names. The service is a combination of freely available national gazetteers and data sets,⁷ as well as volunteered geographic information that is constructed such that, in theory, anybody with an Internet connection can enter or edit data. We obtained data from the gazetteer’s data dump,⁸ and figures 5.6 and 5.7 illustrate the density of place names listed in the gazetteer. In this map, darker shades indicate higher numbers of place names per square kilometer.

The geography of GeoNames is not a simple reflection of the distribution of population. For instance, the United States accounts for 25 percent of the entire database (Graham and De Sabbata 2015), which means that there is more content about the United States than in all of Asia, which contains only 23 percent of place names, and Europe, with 19 percent of place names. India is the most underrepresented country in the world, with only 0.6 percent of the collection (figure 5.7 shows the relationship between population and geographic content in more detail).

Intriguingly, the standard patterns of concentration in the Global North and absence in the Global South are also accompanied by some less expected patterns. Nepal, for instance, has more content than all but ten other countries and contains more place names than India and the United Kingdom combined. We suspect that this is tied to the significant efforts of a 2001 project funded by the European Union (Budhathoki and Chhatkuli 2004) to create a geographical information infrastructure in Nepal. Also noteworthy is that Iran is augmented by almost as many place names as Germany, and North Korea and Sri Lanka are referenced in almost as much detail as Austria. It is likely that some of these patterns appear because many of the place names for locations outside North America and Europe are sourced from the National Geospatial-Intelligence Agency in the United States, and therefore reflect the interests of the US intelligence services. We also note that there are indications that a significant amount of content has been created by crisis mappers (Zook et al. 2010). Haiti, for instance, is annotated by more content than Denmark, which likely is a result of work done following the 2010 earthquake in the country (for more details about the efforts of crisis mappers in Haiti, see Meier 2015).

Because GeoNames is now an essential component of many contemporary digital services (e.g., the geocoding of social media), the presences and absences outlined here have the potential to have a significant impact on how we understand, interact with, and use other digital information. GeoNames may seem like a small corner of the web, but the imbalances noted here can have large reverberations throughout broader information ecosystems.

Similar to the section “Geographies of Participation,” earlier in this chapter, this section has shown that there are broad patterns of unevenness throughout platforms that facilitate the representation of places. Although each platform is characterized by its own idiosyncrasies, some parts of the world are massively overrepresented (i.e., North America and Europe), while the rest of the world is severely underrepresented.

This chapter offers a review of key geographies of access, participation, and representation with respect to open information, using a combination of existing statistics and bespoke data. Through this effort, it demonstrates that in addition to uneven geographies of access to contemporary modes of communication, uneven geographies of participation and representation in open information are evident; and in some cases, they are being amplified rather than alleviated.

Too much should not be read into the specific differences between the geographies of participation and representation. However, apart from the Global South having low scores on every metric, it is noteworthy that rates of representation of Africa and South America are generally higher than rates of participation from Africa and South America. This is because, as noted in “Wikipedia Contributions,” nonlocals can often choose to get involved in the representation of far-off places.

Moreover, there are few signs that global informational peripheries are achieving comparable levels of participation or representation with traditional information cores, despite the hopes of policymakers and the promises of openness. Open (i.e., freely shared information) information geographies seem to follow a pattern similar to other information geographies such as book publishing or academic journal article production; that is, the Global North produces, and is the subject of, exponentially more content than the Global South⁹ (Chan and Gray 2014; cf Dicken 2010; Graham et al. 2011). Even the fast-paced spread of the Internet to 3.5 billion people worldwide has done little to change this basic pattern.

This matters, even though the platforms analyzed in this chapter are all open platforms in principle, because significant barriers remain for people and places in the Global South. As this discussion has shown, these barriers cannot simply be explained by lack of connectivity. As such, any research on obstacles to a broad base of participation in the creation and meaningful use of open information must also seek to consider the impacts of postcolonial modes of informational governance; the culturally contingent ways through which participation and representation take place; shifts in the political economy of control of information (Leszczynski 2012; see also chapter 6); the affordances of platforms and infrastructures of connectivity (see chapter 8 for more on this topic); the constraints of explicit and implicit censorship; gendered or socioeconomic constraints to access or production (see chapters 4 and 9 for more information); and the availability or absence of a broader information ecosystem of local content.

This chapter is a first but important step in showing that global uneven geographies of open information are not simply an outcome of the idiosyncrasies in particular platforms. It also contributes to the literature illustrating how openness itself is always structured and socially embedded (Smith and Reilly 2013), and consequently, it could work to deepen existing exclusions rather than expanding inclusion (e.g., Graham et al. 2016). Openness, therefore, can mean very different things in different places, and it is important to look beyond the term to see what different platforms make visible, invisible, possible, and probable, not just in theory but in practice.

Despite this rather gloomy summation, we see this work as a beginning rather than an end, and also as part of an effort to confront these stubborn realities. This chapter offers an initial broad-scale survey of the contemporary geographies of the production and representation of open information. Future work needs to seek to better understand the contextual causes (using inferential statistics and in-depth qualitative research) and to better explicate some of the contextual nuances (i.e., why certain types of information on certain platforms are more or less likely to be produced in certain places) of these uneven geographies.

This work was made possible by research grants from the International Development Research Centre (Canada), the Philip Leverhulme Fund (PLP-2016–155), and the European Research Council under the European Union’s Seventh Framework Programme for Research and Technological Development (FP/2007–2013)/ERC Grant Agreement no. 335716. The authors are also grateful for the help of numerous colleagues and collaborators. In particular, we would like to thank Matthew Zook, Bernie Hogan, Ralph Straumann, and Ahmed Medhat.

New forms of networked participation and open access in education and scholarship offer great opportunities to expand learning and growth. At the same time, they must flourish and compete in a world of great inequality that is fraught with contradictions. Open access in itself cannot solve the issues of unequal representation in science and scholarship; nor can it solve the tensions between the market and the commons that are inherent in the fluid and unstable global knowledge ecosystem, both of which are compounded by divisions between the Global South and Global North. In fact, this chapter argues that some innovations in openness, such as open access to scholarly publishing, can actually exacerbate those very tensions and inequalities in the system.

The geopolitics of knowledge are characterized by deep fissures and inequalities between the Global South and Global North. These are caused partially by differential resources, but also by historical power relations and determinants of legitimacy. In the academic domain, knowledge patterns reflect physically based geopolitical realities where knowledge from the Global South is treated as peripheral while knowledge from the Global North maintains dominance in terms of all the conventional metrics (Beigel 2014; Czerniewicz, Goodier, and Morrell 2016; Canagarajah 2002; Florida 2005; King 2004). The patent and citation maps produced by Florida (2005) highlight this relationship, showing peaks of these measures of innovation and knowledge clustered almost exclusively in the Global North, with the vast majority clustered in the United States and Europe. The one significant shift since Florida’s 2005 description has been the dramatic rise of China in the citation rankings (Jia 2017). Aside from China’s rise, however, northern dominance remains. At the same time, these knowledge realities are vigorously contested, and also powerfully fueled in higher education today by networked technologies (e.g., Internet and mobiles) in research and education, as well as new forms of scholarship, with new ecologies of access coming into existence.

Widening income disparity around the world, according to the World Economic Forum (2013), is a top global trend and a growing concern. In higher education, inequality is exacerbated by reduced state spending on higher education. For example, between 2005 and 2011, two-thirds of countries who were member countries of the Organisation for Economic Co-operation and Development (OECD) decreased their proportion of public expenditure devoted to education (OECD 2014), and more than half of developing countries reduced their spending on education between 2008 and 2012 (Seery and Arendar 2014, 91). Research funding across developed and developing countries is also inequitable. The gross domestic spending on research and development for the United States in 2012, for example, was 2.76 percent of the gross domestic product (GDP), while for South Africa, it was only 0.73 percent of GDP.¹ Declining funding has led to what can only be understood as the ironically termed cost sharing (i.e., sharing the costs of higher education with students and parents, a much criticized phenomenon).

In this fragile terrain, openness (featuring open content and open scholarship in particular) is often promoted as one means to level these inequalities. Yet ironically, these fissures may actually be intensified by the ways that openness is structured and regulated in the knowledge domains. Of interest in this chapter are the consequences of inequality for educational and scholarly resources. The cost of books, for instance, has risen dramatically—starkly so in developing countries when considered as a proportion of income (Liang 2009).² A related concern is that educational resources are generally not easily accessible to those in the Global South (Trotter et al. 2014).

This crisis in scholarly resources is playing out in a broader context of an increasingly commercialized and financialized education. Of particular concern to the access ecosystem are the threats posed by increased intellectual property (IP) rights, privatization, commodification, corporatization, lack of governmental and corporate transparency, and loss of privacy and overall disempowerment (Hess 2013).

It is in this context that digital technologies and openness are often promoted as one means to decrease inequality. The networked age also promises global digital cultures with flattened power relations, given the affordances of information and communication technologies (ICTs) to collapse distance, enable easier cross-country collaborations, and create new opportunities for knowledge production and sharing (Castells 1996).

The possibilities of the digital world have fueled demands for openness, transparency, creative contribution, and more egalitarian knowledge production and dissemination policies and practices, coupled with a greater consciousness about how knowledge is produced, by whom, for whom, and for what purpose (Comaroff and Comaroff 2012). New formations that have come into existence make it possible for those previously marginalized by infrastructural and physical constraints to participate in global education and scholarship; indeed, networked environments provide potentially hospitable places for collaboration and communication (Trotter et al. 2014). The counternarrative lies in the affordances of the Internet to change the nature of networks by making them more inclusive and easy to participate in (Castells 1996).

This openness is expressed in changing roles and emergent new practices in higher education for all stakeholders forming part of the access ecosystem. For scholars and students, the confluence of the digital with the globalization of education and the mainstreaming of open access (at least in the Global North) has seen reshaped and digitally mediated scholarly practices emerging. At every stage and level of scholarship, new ways of doing research have merged, and boundaries between the old and the new have become increasingly blurred. Within this shifting terrain, openness has been a central characteristic and an emergent practice. Open practices feature increased visibility, more collaboration, and earlier access to research. Ideas and resources have become easily shareable and more granular. New modes, methods, and genres have come into existence, and audiences have become more diverse. Forms of engagement have become multifaceted, and participation more smoothly enabled.

Publishers are also reimagining their positions and contributions to the ecosystem. Until recently, publishers would have known that their core business was concerned with content and that their business model, entrenched for centuries, was lucrative, clear, and understood by all. Now they are moving into the larger field of learning delivery and support, becoming full-spectrum service providers for classroom learning and research; encroaching on tasks performed by libraries, bookstores, teachers and administrators, and technology providers; and incorporating a variety of other student support services. Increasingly, educational publishers understand their competition not as other publishing companies, but as telecommunications companies, software companies, and information-retrieval providers (Gray and Czerniewicz 2018).

What this means for open access in education and scholarship is that assumptions have been profoundly challenged regarding content, its production, and its role in the access ecosystem. The status quo sees traditional and emergent models coexisting and jostling for dominance and legitimacy. The evolving transition to new services and licensing models is not fully understood in terms of who benefits and whose interests are being served. Also, it is clear that the role of openness in this nascent terrain is not yet settled, given the unresolved tensions in the ecosystem of access.

In this chapter, three tensions within the open access ecosystem are identified and discussed: (1) the inclusive versus exclusive nature of openness, (2) digitally open versus free content, and (3) copyright versus open content. Following that, the argument is made that the IP system itself requires rethinking if openness is to be structured in a manner that produces more equitable outcomes.

Openness is associated with availability, with access, and indeed with abundance. It is widely agreed that with networked technologies, digital affordances, and granular multilayered provision, there has been a shift from a scarcity model of education to an abundance model. This has led to what has become known as the “age of abundance” through digital distribution and an end to content scarcity (Colombani and Videlaine 2013); indeed, the total amount of data on the Internet is estimated to be in excess of 7,900 exabytes (Gantz and Reinsel 2011). As the marginal cost of digital products has decreased to essentially zero, perfect digital copies can be produced at no extra cost, making copying essentially free (as well as extremely easy), with profound implications—and opportunities therefrom. There has simultaneously been an explosion in user-generated content. For example, as of May 2019, there are roughly 500 hours of video uploaded to YouTube every minute (Clement 2019). It is the end of the age of scarcity and the models of scarcity on which traditional access systems are founded. Scarcity models are by definition exclusionary, while abundance models promise inclusion. The question is whether they deliver on that promise.

Networked technologies have made the costs of distribution negligible. The ubiquity of open content now offers much to universities, and this is especially compelling given universities’ otherwise resource-constrained circumstances. At a time when a key question has been what can be handed over to technology to save costs, improve efficiency, create opportunities, and derive other benefits, open education and open content would seem to be promising viable solutions, bringing together a set of values and assumptions about shareable knowledge and enabling technology diffusion (see, e.g., Smith and Casserly 2006; Smith 2013). In addition to these efficiencies, the opportunities align with the missions of public universities to serve the public good and the needs of civil society. Thus, there is a clear alignment with the intentions of the famous Budapest Open Access Initiative (BOAI) and the Cape Town Open Education Declaration (CTOED), which both address the Internet’s potential to be an “unprecedented public good” (BOAI 2002) and serve a “collaborative, interactive culture” (CTOED 2007).

However, abundance is not equally distributed and collaboration is unlikely to be equitably experienced. An overt example of these divisions is in the ways that openness does not necessarily serve scholars and students from the margins, such as from the Global South (see Chan and Gray 2013; Rohs and Ganz 2015). There is nothing intrinsically open about the so-called digital sphere, and, therefore, it cannot be assumed that because it enables equitable openness and abundance, it determines or creates it. Rather, it often seems that because the digital educational ecosystem echoes and feeds into the power dynamics of those in the material sphere, it likely reflects those values and indeed may exacerbate or create new forms of exclusion.

An example helps to illustrate this point. The many open access research policy mandates across Europe and North America have led to increased access and availability for scholarship from the Global North, while actually decreasing the visibility of scholarship from the Global South. Without resources (in the form of infrastructure and expertise), scholars on the margins are not able to curate their scholarship in ways that make it openly visible and discoverable. Increasingly, if it cannot be found online, it does not exist. This invisibility and discrepancy are not trivial, given that what is found online feeds into the formation of knowledge production and therefore shapes what comes to be known everywhere (Czerniewicz and Goodier 2014). While access may appear to be more widespread and models seemingly provide abundance, representation and voice are unevenly constituted in new, digitally afforded ways.

In a climate of abundance, there is a widespread assumption that digital equals open and open equals free. However, it is important to note that digital can also afford closed systems and analog resources (i.e., paper) sometimes can offer affordances that can be more open than a particular piece of digital resource might be. The move to digital content is a move from tangible to intangible, and often from ownership to license and digital rights management (the practice of imposing technological restrictions that control what users can do with digital media). In other words, these new digital systems are not neutral. Whether or not they lead to open or closed content is a function of value-laden decisions made by people who impose specific legal and technological frameworks.

Within these new legal and technological frameworks, a complex ecosystem of access is emerging, which exists on continua from analog to digital and from legal to illegal. Legal and open licensed content is only one option in this new ecology. New cultural practices have arisen, including piracy cultures; for many, these are considered both open and free, despite the fact that they are not legally open. Castells and Cardoso (2013, 6) have noted that “All over the world, we are witnessing a growing number of people building media relationships outside [of] institutionalized sets of rules,” and that “a very significant proportion of the population is building its mediation through alternative channels of obtaining content.” They also observed that “piracy cultures have become part of our everyday life in the network society, sometimes even without us fully acknowledging them as such,” and they go so far as to say that “the pirates are more often than not all of us” (Castells and Cardoso 2013, 6). The culture has also been described as an “affective economy,” which introduces an emotional aspect: as has been observed, “filesharing has emerged as one of the most popular dimensions of the new affective economy, in which it simply feels good to share and then it feels even better to embellish, remix, and share again” (Fleming 2012, 685).

There has been what might be called a quiet encroachment. Informal practices are becoming mainstream and acceptable as the blurring of boundaries and the breaking of rules have become the norm. These informal access practices constitute a global phenomenon that are not only limited to the developing world,³ where common misconceptions imagine them to be neatly aligned to deficits of availability. Thus, a narrow view of openness, limited to copyrighted or legally licensed, is not an accurate reflection of the practices and understandings, even by scholars and students (Smith and Seward 2017). This conflation of digitally and freely available with legal is a central tension in nascent ecologies.

Research undertaken under the auspices of research on open educational resources for development (ROER4D) illustrates this tension (see chapter 12 of this volume). It revealed that educators are often unaware of the difference between an openly licensed open educational resource (OER) that can be reused legally and copyrighted material that can be downloaded from the Internet. Educators often assume that online resources are legally available for their educational use (fair use, in legal terms). Less than a third of university lecturers surveyed across nine countries in the Global South said that they had shared their teaching materials with an open license.

Similarly, another study, of 1,000 university students in South Africa, showed a blurring in their understanding of legal and illegal practices and a view that copyright should serve educational imperatives rather than the other way round. Students distinguished between plagiarism, with the author’s right to be attributed, and copyright, which they associated with monetary gains that might be made at the expense of their right to education. The students argued that access to education must be the primary right (Czerniewicz 2017).

In an era where open generally means removing barriers, the legal sphere is one of the few areas where barriers are being added and where the terrain is becoming more complex. Legally, open means specifying permissions, adding layers of bureaucracy, and ironically, perhaps engaging more closely with copyright, rather than less so. This is happening at a time when the general discourse around copyright is about serious criminal actions—piracy, illegal downloading, and file sharing. Copyright is fraught, disabling, enclosing, and bureaucratic. Making copyright better through permissions processes may inadvertently make it more opaque and inaccessible, adding layers of bureaucracy instead of easing access.

This increased bureaucracy is even more problematic given that copyright is now legally the default, as noted by Liang (2005, 105): “Initially, the practices of people operated on the presumption that everything was in the public domain, except where otherwise stated, and copyright did not play much of a role. The history of copyright has centred on a reversal of this presumption to the extent that everything is assumed to be protected unless specifically stated to be in the public domain.”

Thus, openness as a legal solution to the current copyright regime works to legitimize this underlying assumption of automatic copyright—and also undermines the public domain. This adds layers of complexity to a system that previously was more straightforward and presumed openness unless otherwise stated.

Given the tensions noted here, it is useful to consider that copyright in its current form may be the problem and requires rethinking. In the digital world, copying is such an essential action, and so bound up with the way that computers work, that control of copying provides, in the view of some, unexpectedly broad powers—in fact, considerably beyond those intended by copyright law (National Research Board, Computer Science and Telecommunications Board, and Committee on Intellectual Property Rights in the Emerging Information Infrastructure 2000).

It is worth looking at the original intentions of copyright—namely, as a necessary evil for the greater good that should be kept to as short a time span as possible. Baron Thomas B. Macaulay noted in 1897, “Copyright is monopoly, and produces all the effects which the general voice of mankind attributes to monopoly. … It is good that authors should be remunerated; and the least exceptionable way of remunerating them is by a monopoly. Yet monopoly is an evil. For the sake of the good we must submit to the evil; but the evil ought not to last a day longer than is necessary for the purpose of securing the good” (Young 1952).

The fostering of a rich and diverse public domain has always been one of the principal rationales for copyright—“the public domain is not an unintended byproduct, or ‘graveyard’ of copyrighted works, but its very goal” (Birnhack 2006, 60). The intention always was to keep works in copyright for as short a time as possible—which is notably of benefit to the producer, not the intermediary—and works always were assumed to be in the public domain unless they specifically were not. The battle for the public domain remains woefully current, as exemplified by a number of recent disputes over copyright that continue to make headlines—for instance, concerning the works of Sir Arthur Conan Doyle and the “Happy Birthday to You” song.

It is hardly a new argument that knowledge should be free—“as free and general as air or water,” as argued by Lord Camden in 1774, during the court case Donaldson v. Becket—though it has become lost in the current era of commercialization. Goethe proclaimed 400 years ago that “everything that I have seen, heard, and observed I have collected and exploited. My works have been nourished by countless different individuals. … My work is the work of a collective being that bears the name of Goethe” (Frosio 2017). It is a truism to say that knowledge is an accretion, enabled by openness; it was Sir Isaac Newton, in 1676, who famously repeated the maxim of “standing on the shoulders of Giants” [as cited in Turnbull (1959, 416)]. Indeed, scholars cannot produce knowledge without access to other knowledge.

There is presently a serious risk that knowledge, which we could argue based on historical precedent ought to be free and open, is in the “vice grip of commerce,” as “we are in the midst of an enclosure movement in our information environment” (Benkler 1999). Accordingly, it “is caused by the conflicts and contradictions between IP laws and the expanded capacities of new technologies. It leads to speculation that the records of scholarly communication, the foundations of an informed, democratic society, may be at risk” (Hess and Ostrom 2003, 112).

There is a need to rethink the policy environment. The IP frameworks that shape higher education’s engagement with knowledge are anachronistic and outdated, and they are out of sync with the urgent needs of a digitally mediated and extremely unequal world. As discussed previously, it is in part the current IP regime that is shaping the access ecosystem in a manner that tilts the benefits toward those who already hold positions of relative power and voice.

The situation is in danger of getting worse. Many multinational trade agreement negotiations have been vehemently criticized for lack of transparency and the over-reaching provisions on IP. The final version of the IP chapter of the now-defunct Trans-Pacific Partnership (TPP), for example, presented these terms: life plus 70-year copyright term (up from 50); toughened rules against circumventing digital rights management (DRM); targeting whistleblowers and journalists with criminal penalties for accessing secrets online; greater liability on Internet intermediaries; and adopting heavier criminal sanctions. Although the TPP, absent the United States, has now become the Comprehensive and Progressive Trans-Pacific Partnership at the time of writing, these proposed rules show the trajectory of IP negotiations. All these proposals are of serious concern.

How can creativity, scholarship, and learning—elements that are protected in the foundational principles of copyright—flourish in an environment of increasing control? There is a growing need for sustainable IP balances, a landscape of privacy with lots of public spaces, and sustainable development for creativity and public access for everyone to use (Aoki, Boyle, and Jenkins 2006). What happens in the copyright IP arena matters profoundly to education, and new kinds of alliances between stakeholders (including educators and public interest lawyers) are essential. At the very least, institutional IP policies need updating to reflect the changing digital environment, and academics and educators should be taking their IP contracts seriously.

The tensions regarding openness in shifting access ecosystems that are explored in this chapter are being played out and disputed through fundamental contestations in education and scholarship. At the heart of these dilemmas, and thus in the tensions regarding openness, lie disagreements about the purpose of knowledge itself and the interests it serves. The tensions point to the imperative to reclaim the commons and knowledge as a public good. Here, the commons refers to people working together in self-governing groups that govern themselves independent of the state in order to solve problems and share knowledge and resources. While the prevailing dominant imaginary in today’s information societies is market led, the alternative imaginaries are best described as open or commons-led (Mansell 2013). Of course, the current dominance of the copyright regime in its current form is one manifestation of the market-led imaginary. Thus, one step to achieving an alternative imaginary is the rethinking of copyright as discussed earlier in this chapter. But there is more to do.

For those wishing to ensure that knowledge remains free, with education that is fully inclusive and scholarship for the purview of all, increased attention needs to be paid to the nature and governance of the commons (specifically the digital commons). The commons are institutions that fall between markets and states and are vulnerable to social dilemmas and threats of enclosure (Hess 2013). It is an often-used term, with the knowledge commons or cultural commons forming only one part. Originally, it was applied to the natural commons, and it is important not to conflate the two. In the natural commons, physical resources can be overused, and one encounters depletable and exhaustible participation, rivalry, and scarcity—the so-called tragedy of the commons, wherein individuals acting in their own self-interest deplete the shared commons (Hardin 1968, 1244). In this context, open access has very different and negative connotations.

The digital knowledge commons, by contrast, is physical but also intangible, generative and regenerative, wherein peer production is nonrivalrous and offers abundance. Open access in this context is a very good thing, and, if anything, there is a threat of underuse (Hess 2013). The commons for education places knowledge producers in charge. Here, cocreation and ethical and social benefits become the norm, and the knowledge economy shifts to a knowledge society. This is not merely a semantic shift; it is actually a profound conceptual shift.

Academic scholarship is premised on collaboration and sharing, and it is an educator’s prerogative to reclaim knowledge as a nonrivalrous and nonexclusive public good. Individuals, alone, are not enough to stake this territory. If support for OER does not come from public bodies, such as governments and universities, the OER movement could lead to control of the system residing within a technological elite, a narrow band of activist producers, and a corporate takeover of the business side (Jones 2015).

For equitable openness to be achieved, the state has an important role to play, although not in organizing and managing education in a homogenous way. The state needs to support the 2012 Paris OER Declaration of the United Nations Educational, Scientific, and Cultural Organization (UNESCO) and policy frameworks to enable open education, to invest in the knowledge commons, and to steer the market through light touch interventions and disincentives. Universities play an essential role in owning, controlling, and managing knowledge production and dissemination, and in a digital age, they should be asserting their rights to manage their own assets. They should assert the priority of academics and authors as the agents and owners of knowledge, protect the autonomy at the heart of commons-based structures for knowledge, and develop and support collaborative initiatives in knowledge dissemination. Overall, open education is a means to an end, and it is only one—albeit an important—strategy toward an equitable, democratic, and peaceful world.

Grateful thanks for the rich resources provided by Giancarlo Frosio LLM, whose presentation (Frosio 2017) provided links to valuable references used in this chapter.

This chapter starts with the premise that the impacts of open activities on economic and human development are large and increasing. These impacts come through a variety of mechanisms and are affecting all sectors. Governments are starting to recognize the opportunities presented thereby and are enacting related policies. As of early 2020, seventy-eight countries have made commitments to “make their governments more open and accountable”¹ (see chapters 10 and 11 for more on open government activities). Many countries are enacting policies in support of open educational resources (OER), with the hope of improving quality while reducing costs for education (see chapter 12 for more on OER). Nongovernmental organizations (NGOs) and community organizations are taking advantage of free software tools, content, and the ease of connecting with others. To wit, around the world, there have been over 90,000 deployments of Ushahidi, an open-source crowdsourcing platform, to address local issues.² Private-sector companies are similarly extracting economic value through new open business models. Studies by the Computer and Communications Industry Association have estimated the share of open activities in economic production in the United States to be about one-sixth of gross domestic product (GDP) in 2010 (Benkler 2011, 314–315), and it is growing fast.³

In this chapter, we focus specifically on how these changes, brought about through various open practices, might necessitate new economic understandings and explanations, specifically concerning the role of economic policy and theory in international development. The perspective is primarily micro, with some more macro dimensions of open provision explored later.

The starting point for this exploration is open provision; that is, the provision of the innumerable goods, services, and benefits that the private, public, and nonprofit sectors make available to consumers and recipients free of direct charge or price. The viability of open provision is based on leveraging open informational resources and voluntary human resources at little or no direct cost, together with—in the private sector in particular—revenue streams from many kinds of marketing, as well as selling consumer and/or market data to suppliers. Its expansion has been very innovative with respect to inventing and applying new provision processes and institutions. As we will see in this chapter, open provision is happening in the context of the rapidly expanding number and form of open practices throughout economic, political, social, and cultural endeavors.

This chapter begins by further defining and providing examples of open provision, and then briefly sketches two main frameworks to examine the economic dimensions of open activities: normative welfare economics, which focus on the provision of goods and services (henceforth referred to as goods); and the human development and capability approach (HDCA), which focuses on the expansion of human capabilities and freedoms (economic, political, social, cultural, and ethical). Following this is a policy-oriented exploration of the relationships among private, public, and open goods. This exploration reveals, in particular, that open provision extends (or, with negative goods, diminishes) both public and private goods and freedoms, thus changing the role of markets, nonprofits, and governments in providing them. Finally, the chapter provides a number of perspectives on policy and practice, with the objective of expanding the roles of open activities in the provision of goods and the advancement of human well-being.

Openness has many uses and meanings (Lundgren and Westlund 2016; Pomerantz and Peek 2016). Here, we draw on Smith and Seward (2017), who argue that openness is better understood in terms of processes and practices (e.g., sharing), rather than artifacts (e.g., open goods or services). They identify three main types of open processes: open production, open distribution, and open consumption. Each process consists of a series of open practices that drive the process. A quick overview of open processes and practices, with examples from the private, nonprofit, and public sectors, is provided in table 7.1.

If the provision of goods amounts to their production, distribution, and consumption, we define open provision to be the provision of goods and freedoms that include open practices in one or more of the provision components (production, distribution, or consumption). The examples provided in this chapter indicate that few (if any) goods and freedoms are completely open; at a minimum, most have important information and communication elements that use information and communication technology (ICT) infrastructure and services provided by nonopen proprietary market practices.

In terms of artifacts, open goods and open freedoms have open practices prominently in their provision components. But as noted, we prefer a process view of provision due to its accuracy and clarity. In this view, open provision is the provision of goods and freedoms that include open practices in one or more of its components. The provision components—production, distribution, and consumption—are further disaggregated in the following examples. The principal open practices are peer production, crowdsourcing, sharing/republishing, and retaining/reusing/revising/remixing (the 4 Rs). In addition, nonproprietary is a characteristic common to open practices, and free to consumers is typical of goods with open practices prominent in their provision.

In this section, we present a few types of open provision that incorporate the three open practices of sharing, peer production, and crowdsourcing. All of the examples presented here are important in both less and more developed economies and societies, with some having particular importance within developing countries. It is important to note the following:

•These examples cover the full provision process (input provision, production, distribution, and consumption) of the good or freedom.

•These are stylized provision business models that do not include any focus on revenue and viability issues.

•These are types rather than specific cases (e.g., open-source software versus Mozilla Firefox or Open Office).

•These tables leave out intermediate inputs and focus on the provision of value-added and end-use consumption. Intermediate inputs typically include open activities in their provision; as a result, there are few (if any) goods and freedoms that are completely nonopen.

While we provide five case types in the following discussion, we have outlined provision models for many other specific cases. The striking feature of these instances is not that they fit easily into specific typologies, but in fact that they are dramatically diverse. The extent of innovation and adaptation of open practices in provision of goods and freedoms is truly remarkable.

Examples of open peer production provision (e.g., open-source software) abound and include the popular web browser Mozilla Firefox and the Apache web server that, as of April 2017, was estimated to serve about half of all the existing websites.⁴ The Open Medical Records System (OpenMRS) is a collaboratively developed, open-source electronic medical record system platform that was developed to work in resource-constrained environments.

When open-source code is developed and maintained collaboratively, open-source software provision entails open practices throughout the provision process, with market components mainly being in the ICT platforms used. However, like many other open informational resources, there are also private-sector companies that make their code open source for a variety of reasons.

This commons-based peer production model is also applied in areas outside software development, typically with the same set of open practices throughout the provision cycle. For example, India’s Open Source Drug Discovery project hosts an open and collaborative research platform “to accelerate drug development for neglected diseases” (Masum et al. 2013, 113). Wikipedia, one of the most highly accessed and used websites in the world, is also a variant of the peer production model. Similarly, iFixIt is a wiki-based, privately run website with collaboratively produced do-it-yourself (DIY) repair manuals for devices. Their business model includes selling the parts and tools required.

A critical element of peer production is that the open practices it entails can spawn cycles of knowledge growth, reuse, and sharing, which lead to increasing returns and knowledge spillovers (Garzarelli et al. 2008).

Ushahidi is an open-source crowdsourcing platform developed in Kenya (and is now also offered as a software service by a private software company).⁵ It has been locally implemented tens of thousands of times in both developed- and developing-country contexts for a variety of purposes, such as security, transparency, and crisis survival and management. One Ushahidi implementation, HarassMap,⁶ is an Egyptian crowdmapping site for reports and testimonials of sexual harassment. The NGO that implemented HarassMap goes beyond the collection of knowledge voluntarily from participants (crowdsourcing) and leverages the open sharing of data as part of a larger process of social mobilization and sensitization, as well as change.

The public sector can also engage in citizensourcing, the crowdsourcing of data by soliciting information from citizens. Citizensourcing is one of a larger set of open government practices that taps into the collective knowledge of the public to produce public value creation (Hilgers and Ihl 2010).

Similar examples, but with different social and economic implications, are private sector–based crowdsourcing initiatives that then share the gathered information as part of a service. For example, TripAdvisor is a travel information service that can be accessed via either a website or a mobile app.⁷ The principal data for this service come by crowdsourcing the knowledge of travelers in their reviews of places, hotels, restaurants, and other travel businesses, freely provided and shared with everyone. The rest of the provision activities are market provided (see table 7.4). Revenue to the site appears to come primarily from advertising and the selling of user data to travel and other service providers. Many similar private services, such as Yelp, also hinge on views, reviews, and evaluations by consumers and users.⁸

These services are expanding in every subsector of the economy. In finance and investment, for example, direct trading platforms provide free market data and research to investors, deriving revenue from large volumes of trades and sale of market activity and user data. In real estate, sites increasingly provide free data (e.g., assessments) and links to service providers (e.g., financial, legal, or staging).

In this section, we will examine two examples: OER and open government data. OER are defined as “teaching, learning, and research resources that reside in the public domain or have been released under an intellectual property (IP) license that permits free use and re-purposing by others” (Smith and Casserly 2006, 9). The first prominent example of OER came in the form of the Open CourseWare (OCW) offered by the Massachusetts Institute of Technology (MIT), where MIT made the content of their courses openly available on their website in 2002.⁹ Since then, OER have taken thousands of forms, from open textbooks to online course modules to massive open online courses (MOOCs), with important implications for the societies and governments of countries in the Global South (Smith 2013).

There are a huge variety of OER available and a very diverse set of production models for their creation. For example, Siyavula Education, an education technology company in South Africa, emerged out of a group that facilitated a collaborative authoring process of openly licensed textbooks. Open textbooks are freely available to teachers and students across the country to download and use as desired. Another model of open production was tried in the state of Utah, when the government decided to support open textbooks in key curriculum areas in 2012. A pilot study found that this model of producing open textbooks can reduce costs by 50 percent or more (Wiley et al. 2012).

OER have many users, and educational initiatives can leverage the existence of a wide variety of high-quality OER. For example, the Darakht-e Danesh project, run by the nonprofit group Canadian Women for Women in Afghanistan, created a web portal for open educational content from around the world, and curated and translated to fit the Afghanistan curriculum (Oates and Hashimi 2016). Afghanistan’s teachers can and do contribute content to the web portal as well. Another example is the Rumie Initiative, run by a nonprofit organization based in Canada. Experts and communities choose top-quality open educational materials, put them on tablets, and provide them to underresourced schools in developing countries and to First Nations communities in Canada. The principal inputs for the Rumie Initiative are OER, ICT connectivity, very low-cost tablets produced by a supplier in China, financing (partly crowdsourced online), and management (a virtual network of individuals and organizations).

Open government data are publicly shared governmental informational resources. Governments typically disclose their data for a variety of reasons that “run on a spectrum between service delivery and public accountability” (Yu and Robinson 2012, 182) ideally resulting in a wide range of potential economic, social, and political benefits (Verhulst and Young 2016). One example of a public accountability application of open data in Uruguay is A Tu Servicio, an app that provides an easy-to-use interface to open government health data, allowing citizens to compare local healthcare providers based on a variety of key indicators (A Tu Servicio is discussed in more detail in chapter 11). The traditional provision cycle of open government data contains only a few open practices at the delivery and consumption end (see table 7.6).

There are several other models of open government data provision beyond governments making existing data available. Governments and NGOs, for example, run crowdsourcing processes to generate relevant data. Assessments of when such ICT-enabled citizen voices lead to government responsiveness is provided by Peixoto and Fox (2016). It is also common for intermediaries such as NGOs to work with open government data, often repackaging and republishing the information for a particular purpose and audience (van Schalkwyk et al. 2014). In other instances, NGOs might gather government-held data through a variety of means (such as freedom of information requests), compile and curate them, and then share them publicly as open government data. For example, in Mexico, the Mejora Tu Escuela (Improve Your School) project works to encourage parents’ engagement in the education of their children and to combat corruption in the education sector (Verhuslt and Young 2016). It does so through a portal that presents (mostly previously inaccessible) data from Mexico’s Ministry of Education, including information on school infrastructure, locations, standardized test results, and teacher assessments. The portal also includes some national census data, which were not originally released by the government but were leaked to NGOs (Verhulst and Young 2016).

Key parts of the open provision ecosystem are the platforms that enable open practices and the provision of goods and freedoms, which themselves have varying provision models. The platforms can be either technological infrastructure (such as the Internet) or the application layer (typically websites or apps). The Internet is mainly privately provided and not free to users. The major access and social platforms, including those of Google and Apple as well as Facebook and Twitter, are generally free to users. Their services do come at some cost to the user in terms of privacy and loss of control over the data that they generate from using the platforms; indeed, the huge number of users makes the platforms highly valuable for advertising, as well as consumer and market data acquisition.

Sharing digital informational resources at scale, typically over the Internet, has emerged as a central practice since the middle of the 1990s. The rise of the prevalence of open processes is directly linked to the increasing spread of low-cost connectivity and information capabilities of the digital revolution, in the form of increasingly accessible and less costly Internet and information technology platforms and means.

This rise of open processes and provision in all institutional economic sectors—public, private, nonprofit, and personal (individual and household)—has meant that the economic impacts of open provision have grown considerably in the last two decades (Benkler 2013). This economic activity emerges not just from new models of peer-to-peer production that fall outside of the public and private sectors, as Benkler describes, but also from state (e.g., open government data) and private actors.

The reason that open practices are gaining in significance is because they can add value and reduce cost (i.e., increase efficiency in the provision of goods and freedoms) or provide novel and more effective means to accomplish specific goals or fill particular demands (see table 7.7 for a brief overview of the benefits and costs of engaging in the various open practices). Of course, there is no such thing as a free lunch—content might be free (in terms of price), but there are always some associated costs, be they connecting to the Internet or other resource costs or abdicating control over personal data and privacy. These costs need to be covered in some manner in order for provision to be viable. Open practices in provision also often help to reduce information asymmetries between providers and users, thus helping with another aspect of inefficiency in market provision.

Many of the kinds of benefits and costs in table 7.7 are very difficult, if not impossible, to measure. This is because open activities typically are free to users, and their inputs are often voluntary—that is, there is no wage or price attached to them or they are difficult or impossible to value for other reasons. Consequently, their value does not receive proper consideration or weight in national accounts and productivity estimates.

Economic analysis of open provision needs a particular economic framework. Here, we draw on what we consider to be the two most relevant frameworks: normative welfare economics (market economics plus public finance) and the human development and capability approach (HCDA). The former, which focuses on the provision of goods and services, has been at the heart of the postwar Western economies and globalization. The latter is, in fact, much broader than an economic framework per se, as it focuses on the provision of capabilities and freedoms—economic, political, social, cultural, and ethical.

Normative welfare economics, as described by Smith et al. (2011, 83n13), are observed to be a “body of economic theory that addresses maximizing the material welfare of a society and its individuals—welfare derived from consumption of goods and services—and public finance for public goods and services. … Normative welfare economics includes equity as well as efficiency considerations, and provides analysis of equity characteristics of market functioning and of public economic activity—expenditures, taxes, policies, and regulations.”

Normative welfare economics recognize that governments play a key role in the provision of public goods such as national defense, health, education, law/justice, environmental protection, and social services. These goods are either not typically organized by market mechanisms (defense and justice) or privately provided in suboptimal amounts (health and education). Public goods comprise about a quarter of the GDP of most economies, and, next, we explore their characteristics in detail in comparison with those of open goods.

Our second economic framework is the HDCA. A detailed introduction of this concept is beyond the scope of this chapter, but its central ideas include human functionings, capabilities, freedoms, and agency.

The definitions of the key terms of the HDCA are as follows:

•Functionings, as noted by Sen (1999a, 75), are defined as “the various things a person may value doing or being.” In other words, functionings are valuable activities and states that make up people’s well-being, such as being healthy and well nourished, being safe, being educated, having a good job, and being able to visit loved ones. They are also related to goods and income, but they describe what a person is able to do or be with these. For example, when people’s basic need for food (a commodity) is met, they enjoy the functioning of being well nourished.

•Capability refers to the freedom to enjoy various functionings. In particular, the term is defined, by Sen (1992, 40), as “the various combinations of functionings (beings and doings) that the person can achieve … reflecting the person’s freedom to lead one type of life or another.” In other words, capabilities are “the substantive freedoms [that a person] enjoys to lead the kind of life he or she has reason to value” (Sen 1999a, 87).

•Agency, in turn, is defined as the ability to pursue goals that one values and has reason to value. An agent, therefore, is “someone who acts and brings about change” (Sen 1999a, 19).

Perhaps the key area of difference between normative welfare economics and the HDCA framework is that the latter focuses on more dimensions of human prosperity or flourishing—specifically, social, political, cultural, and ethical dimensions—than just material (i.e., economic) ones (Sen 1999a). Table 7.8 provides an overview of the differences between normative welfare economics and the HDCA.

Welfare economics draws a blank on what specific public goods should be produced and consumed because consumers’ preferences do not translate directly into production decisions, as with markets and private goods. One central pursuit of the HDCA, therefore, has been to develop workable theories and processes of social choice and justice (Sen 1999b). In this respect as well, work that has extended the original HDCA conceptualization recognizes and emphasizes the central role of power—and the interests of the powerful—in governance and policymaking (Spence and Deneulin 2010). This is a dimension often absent in the analytic focus of normative welfare economics on efficiency and equity.

Open informational resources, practices, and provisions appear to have strong implications, in terms of both potential and challenges, for equity, informed public discourse, social and public choice, and the provision of all varieties of freedoms.

Many meanings of the term public goods are in current use, based on attributed characteristics of the goods (artifacts). We review these perspectives here before suggesting better process (provision) views or definitions of public goods.

As previously mentioned, normative welfare economics considers public goods as those that the market or the nonprofit sector would not provide in the right amounts (or forms), and thus their provision is organized and also might be carried out by government. In this conception, public goods, in fact, are mostly services. Government provides physical goods, such as roads and military equipment, but most government economic involvement is by providing services, including organizing the provision of services, financing them, and actually producing and delivering them. Government organizes and finances defense, for example, and military forces (which are essentially governmental) deliver the service, while private industry provides military hardware and much of the technology. In many countries, sectors such as education and health are both publicly and privately provided. Government is also involved in regulation in all sectors, and regulatory services are particularly prominent in some.

Public goods have also been characterized in terms of two main aspects—excludability (whether individuals can be excluded from use of the good) and rivalry (whether one individual’s use of a good reduces its availability to others). These concepts have been key to arguments about why their provision is not done at all, or not done well, by markets. A well-known matrix that characterizes (material) goods and services is shown in table 7.9.

We suggest that this characterization of public and private goods requires some updating. The world has changed considerably since this original formulation, in particular with the advent of increasingly abundant free and open informational resources and the emergence of open practices and provision. Hence, it would be informative to bring this table up to date and include newer examples of goods and services that contain open processes. These examples challenge the idea that goods are rivalrous or excludable by nature, regardless of how the component activities of their provision are organized. They also help to explain the underlying reasons for rivalry and exclusion, which look particularly important in the current world of digital content. In this section, we offer an alternative to the rivalrous or excludable matrix to understand public and private goods in relation to open goods.

The notion behind excludable goods, which include doughnuts and journal subscriptions, is that it is possible to restrict who can consume them. Because each individual receives all the benefit, sellers can appropriate the revenue, and these goods are generally provided and sold by private market organizations. Nonexclusive goods, such as libraries and free-to-air television, are typically provided by public and nonprofit organizations (although sometimes by the private sector as well) free of direct charge (i.e., price) to consumers.

However, the distinction between exclusive and nonexclusive goods, as originally made, does not hold as accurately now, particularly with the advent of the Internet and digital goods. In practice, it is perhaps more accurate to assess exclusion versus inclusion on the basis of how the good is provided (process), rather than its inferred nature (artifact). Movies and television programs, for example, are nonexclusively provided on free-to-air television via antennae (of which there are few left), exclusively on cable or paid Internet services (such as Netflix or Hulu), and inclusively on free-to-user Internet services (such as YouTube).

The notion of rivalry is also not as clear cut in a digital world. The basic notion is that consumption or use by one person either prevents or detracts from use by, and benefit to, others. On the side of exclusive goods, this separates out club goods, which are physically consumed jointly by users (cinemas, private parks, or airplane, train, and bus rides) but still are paid separately by each consumer. Some club goods (movies) can be rivalrous or not in different provision models (cinemas, Netflix). Club goods comprise a relatively small subset of exclusive goods—those physically consumed collectively.

On the side of nonexclusive goods, rivalry is said to separate out common goods, such as timber and fish stocks, because their supply is so large relative to demand that they are considered to be effectively free. However, are there many common goods left? Real wilderness areas (not public parks that people pay fees to visit), for example, remain a dwindling luxury in most developed countries, and increasingly so in developing ones. Even clean water and air are increasingly scarce in many cities around the world. Unlike the diminishing physical commons, the knowledge commons is growing rapidly, with growing populations and communications adding to stocks rather than diminishing them. Nonrivalry—one user not diminishing availability to others—is in fact a powerful basis of much of the explosion of the open provision of goods and freedoms.

Given the revisions just indicated, we propose a revised matrix of exclusiveness and openness in goods provision, shown in table 7.10.

In this schema, open goods are found in five categories: open publicly (state) provided goods and services, open private services that are free to users, common goods, knowledge commons, and other largely open goods, including open-source software and OER provided often by the nonprofit or voluntary sector.¹⁰

Many publicly provided services are starting to incorporate open practices in their provision. Indeed, this even means that sometimes the benefits of some government activities, such as funding open textbooks as part of their education provision, can spread beyond the borders of that government’s particular jurisdiction. Further, within the public sector, there are highly open goods and freedoms based on the provision of open informational resources (e.g., open data initiatives and referendums).

How much insight does one gain by categorizing goods in these ways? In normative welfare economics, the defining of public goods was intended to specify which goods are not provided (either efficiently or at all) by markets, yet are demanded by people, so their provision needs to be organized by government. The emergence of open goods adds complications and a more nuanced view of government roles.

More specifically, open provision models arguably offer alternatives where both markets and the public sector fall short of supplying goods and freedoms that are in demand. To market provision, they add new goods (such as free applications and content) and services (such as user reviews) that help to reduce information asymmetries between sellers and buyers. To public provision, they add new ways (such as referendums) to identify demand for public goods and freedoms, greater reach and quality of public services (e.g., through OER or by offering open government data to make public expenditures and activities more transparent), and new goods and freedoms (such as crowdmapping and crisis management) that need voluntary input to succeed.

Although public policy has many purposes in addressing public goods, key questions in both normative welfare economics and the HDCA frameworks is about what public goods to provide and how to organize the provision. With the explosion of new open goods and freedoms now occurring, reformulating the theory of public goods is needed to guide public provision decisions and policy and improve informed public discourse, social choice, and social justice.

In this context, an important feature of open provision in the personal and nonprofit sectors is simply the possibility that people can organize the provision of collective benefit goods and freedoms without the process requirements and constraints of public organization, consensus requirements, and funding. On the whole, this should add greatly to the overall efficiency of providing valued goods and freedoms and, hence, to the flourishing of humanity. However, as chapters 1 and 6 illustrate, open provision can lead to the increased efficiency of provision of social bads as well as goods. This concept will be discussed briefly in the following section.

Phenomena as large as open provision might be expected to have a wide range of major implications for public policy and action. Here, we have picked out some main areas from the analysis—namely, public provision of digital goods and freedoms, IP, employment and social protections, privacy and security, and a cluster that includes agency, social choice, and political freedoms.

To what extent should governments be active in the provision of open goods and freedoms? While the particulars of each national government will be highly influential, we provide next just a quick review of a few areas where governments are engaged or could benefit from engaging in open provision. This brief discussion is indicative rather than comprehensive, focusing on goods provision; public curtailment of the provision of bads is addressed separately in the “Privacy and Security” section later in this chapter.

Education Given the public goods nature of an educated population, it is not surprising that most governments attempt to provide public education for all. However, there is variation in the abilities and commitments of governments to cover the cost of educational resources. Some governments are adopting open practices in the provision of educational resources, such as making all educational resources that are funded by the government openly licensed, and thus reusable for free by students. There are some emerging models that seem to be working. For example, in British Columbia, Canada, the Ministry of Advanced Education launched the BC Open Textbooks project to provide open textbooks in the subject areas with the highest enrollment levels. With this model, the government pays comparatively less than the original outlay to have the books updated annually. This means that not only are the books free to teachers and students, but they are also more up to date than previously published alternatives. Such an approach may be doubly beneficial for developing countries if there is a focus on working with in-country authors, considering that foreign publishers are often highly active in developing-country markets (Toledo et al. 2014).

Governance As discussed previously, open data initiatives by governments have expanded rapidly, with the potential to improve governance across a wide range of areas. The availability of open government data is also argued to enable much value creation by the private sector—estimated by some to be in the trillions worldwide (Manyika et al. 2013). Furthermore, there are also many examples of government use of crowdsourcing to improve service delivery in education, health, water, waste management, and other areas, as well as a means to foster broader public participation (Bott and Young 2012; Gigler and Bailur 2014; Peixoto and Fox 2016). In many countries, both less and more developed, improving the public, private (e.g., social media), and nonprofit mechanisms furthering political engagement and consensus building is a key public policy challenge.

Intellectual Property Protection While IP has a long history, its relative significance for economic and human development is arguably only increasing given our shift toward increasingly knowledge-based societies. IP protection has been supported mainly on the ideas of fairness to inventors and incentive for research and development, and, on the other hand, it has been challenged throughout its history on the grounds that it stifles innovation. The current copyright regime hampers one of the core affordances of the Internet (easily sharing digital goods) and thus limits the sharing of ideas, knowledge, and resources that are fundamental to key social functions such as education (Wiley 2017). The emergence of open provision provides a fundamental challenge to the current dominant IP regime and competition policy. Drawing on examples of the major innovations in the Internet, Yochai Benkler (2011, 318–320) argues that the freedoms extended by more flexible IP policy (“freedom to operate”) are more important than strong IP protection in driving this innovation.

There is also an emerging body of knowledge on practical ways in which knowledge is appropriated and used for innovation in informal sectors of developing countries. In a synthesis of five case studies of informal sector innovation in sub-Saharan Africa, de Beer and Armstrong (2015) found that informal sector innovators adopted open practices (particularly sharing), although they were not necessarily digital. The key point is that the protection of IP rights did not play a significant role in spurring innovation.

Similarly, Mizukami and Lemos (2008) describe a business model for local Technobrega musicians in Brazil that is also predicated on giving away their music to street vendors, who then sell it. The goal for the musicians is to share their music broadly to gain popularity. Then they host street concerts for which they can charge admission. It mirrors online freemium models, which give away some basic services or content, but then charge for extras. These perspectives are particularly significant, given that in many developing countries, the informal economy constitutes the majority of the economic activity.

Much more could be said about IP policy and the law. Open innovation analysis, for example, describes the relationship between open innovation and IP and recommends appropriate IP and other marketplace framework policy measures (de Beer 2016). For example, there is an argument for policy neutrality between the proprietary firm-based system developed throughout the Industrial Revolution and the rapidly emerging open and collaborative needs of open innovation and provision.

Nonneutrality of infrastructure poses another large set of diverse interests facing IP and competition policy (Fuster Morell 2014). Monopolization occurs in many ways through a demand-side discoverability dynamic where, for example, despite the wide diversity of a particular type of content, those elements that are easily found on Google will dominate greatly, while the rest exist in the long shadow of dwindling use (see also chapter 6).

Employment and Social Protection It is difficult to measure the current extent of open provision and how far and fast it will expand. Available studies suggest that open provision—in the sense of industries that depend on and benefit from limitations to copyright, rather than its extension—comprise about one-sixth of the US GDP (Benkler 2011). In the case of open provision, trends or limits might be easier to approach than current levels. One interesting perspective on the potential is the aggregate amount of time that individuals spend on the consumption of open-activity goods and freedoms. According to Shirky (2010, 10), “Americans watch roughly two hundred billion hours of TV every year. … Even tiny subsets of this time are enormous: we spend roughly a hundred million hours every weekend just watching commercials.”

Open provision looks likely to expand greatly on the basis of so-called big data and data exhaust, as well as sheer innovation, rather than the amounts of human and capital resources used in provision. The impact of artificial intelligence (AI), for example, is already substantial and growing fast. One large area includes the studying of user and market behavior and the tailoring of provision and marketing to individual preferences. Where AI may lead is hard to guess. Its possible applications were suggested by Sebastian Thrun, founder of Google X (Thrun 2017):

My students and I recently did work on artificial intelligence for detecting skin cancer, and we found that if we train an artificial intelligence with about 130,000 images, we can find skin cancer basically using an iPhone as accurately as the best board-certified dermatologist. … Every time I talk through my phone—and it’s probably about an hour a day—it could analyze my speech and thereby find things like Alzheimer’s much, much, much earlier than we find it today.

Open provision in many cases displaces jobs and employment (e.g., appliance repair people and real estate agents). AI and robotization extend this potential in both lower-and higher-skilled employment. This is one part of a picture of uncertain employment prospects associated with technology, automation, digitalization, and globalization, among other areas. Employment and social security implications require public policy responses.

It is noted that open provision models, including intelligent job search and matching, counter the “fatalism that we are powerless to harness what we create to improve our lives—and indeed our jobs” (Spence and Manyika 2015, 1). Nevertheless, open provision contributes to the possibility of widespread prosperity without widespread employment and the need for political action and social protection (e.g., guaranteed income) to resolve who gets what. In many developing countries, where social and economic security mechanisms are not very developed and where full-time employment is rarer and temporary or freelance work more common, current and coming challenges are substantial.

Privacy and Security It is clear that open provision can provide bads and restraints, as well as goods and freedoms. One example is the use of online platforms where the ease of sharing can support the recruitment and radicalization of individuals, who in turn may participate in terrorist activities. Similarly, one can use these platforms for bullying, hate campaigns, and political manipulation where provision reduces rather than expands freedoms. Another set of privacy and security concerns surrounds major open platforms and the theft and misuse of personal data from user repositories (cybercrime). These and other concerns (e.g., cyberwarfare) apply to proprietary (corporate, military) provision processes, as well as open ones and, while important, are well beyond the scope of this chapter. One thing that seems clear is that the Internet and the high-volume provision that it enables—both open and proprietary—are facing much higher security and privacy costs than in the past, as both security and privacy are essential to sustain openness and broad public participation. They are also facing major substantive challenges for public policy and participation, reflecting greater difficulty than expected in identifying and managing or countering highly negative behavior.

Agency, Social Choice, and Political Freedoms Open provision can promote agency, social choice, and political freedoms, which are three major concerns of the HDCA. The idea that open provision activities tend to promote agency does not seem farfetched when one looks at people’s personal contributions in terms of commitment, time, effort, skills, knowledge, experience, management, and finance. In many cases, connectivity and platforms allow people to be engaged in the provision of goods and freedoms that they value and have reason to value. Open activities certainly can serve the provision of informed public discourse, social choice, social justice, and political freedoms, which are not small matters. For instance, public consultation processes are examples from the public sector, and many of the Ushahidi-based initiatives are examples on the nonprofit side.

On the other hand, open activities can also pollute public discourse and thus effectively reduce political freedoms and social choice. One prominent example is the emergence of deliberate misinformation as legitimate news stories (i.e., what has become known as fake news) as a genuine threat to democratic governance. The prevalence of misinformation has become significant enough that some social media platforms, such as Twitter, have proposed measures to counter its spread. Security dimensions of fake news have also been highlighted by Russia’s activities before and after the 2016 election in the United States. Given the effectiveness of these techniques, such interference in political processes is increasing—with organized disinformation campaigns found in 48 countries in 2018 (Bradshaw and Howard 2018.)

Furthermore, the tailoring of free content based on an individual’s past actions by social media platforms serves only to further fragment society, creating separate realities and alternative facts, dramatically undermining the possibility for meaningful dialogue. Indeed, it is an open question as to whether democracies will survive in such a context.¹¹

The reality and rapid expansion of open provision necessitate a change in the nature of economic theory and analysis. A simple example from this chapter is that it is no longer adequate to address economic efficiency, equity, and stability in terms of private and public goods. Rather, one must include open provision across the public, private, nonprofit, and personal sectors. Digital provision—both open and proprietary—is transforming economies in terms of goods and freedoms provided, work, IP, competition, and security. Because proprietary provision has been the dominant mode of market economies, the large open portion of digital provision is, in our view, the most transformative part. Like the future of open provision, its ultimate impact on economic thinking is in the early stages and very hard to predict. Nevertheless, the discussion should prove both valuable and interesting.

Open access has had a positive effect on all things digital, including broadband. However, there is limited empirical research on its application as policy in developing countries or its impact or effectiveness in such contexts. Many of the assumptions underlying the theories and practices of open reflect conditions of abundance and freedom that are characteristic of mature, competitive economies and democracies of the Global North (see, e.g., Benkler 2013, vii–ix). As such, open access strategies cannot be applied uncritically to developing countries where the local context may not be conducive for success. There is also limited research on the potential of leveraging private and public resources through open systems to deliver public goods such as broadband, where states may not have the capacity to do so on their own.

This chapter seeks to contribute to a better understanding of open access regulation as a policy instrument in the context of developing countries. Building on the notion of private provisioning of social and public goods (Hirschman 1958), the argument presented in this chapter is that where there are no resources or institutional endowments for the state to provide social and public goods, the only way of doing so (in this case ensuring digital inclusion) may be accomplished by leveraging private resources. Drawing on Frischmann (2005, 2012), this chapter also makes the case for recognizing the often-ignored demand-side value of infrastructure, especially in the case of redressing digital inequality and citizens being unable to pay for services essential to their economic and social participation in modern society (see also Gillwald 2016).

The theories discussed in this chapter are used to create a framework to conceptualize and examine various types of open access in developing countries: (1) the mandatory open access regulatory regimes that emerged with the liberalization of markets and the regulation of former monopolies, in the interests of enabling the interoperability of networks and access to wholesale networks and facilities by service providers; (2) the mandatory and voluntary competitive models that have evolved with new broadband networks required to meet the demand for data with the popularization of the Internet; and (3) the resource management form of open access found in the theory and practice of the commons. In some cases, the latter is a subset of the governance framework described in the first point. This is true, for example, in the unlicensed spectrum set aside for public use that is examined here in the case of free public Wi-Fi as a state intervention to meet the communications needs of its citizens. What can be seen as a commercial variant of this service is also examined here: private operators zero-rate products or services by allocating bandwidth for open use in order to attract customers by providing free access to a product or service.

The chapter proceeds as follows. First, it discusses the evolution of open access regulation. It then extends this conceptualization by discussing the three approaches outlined previously: mandatory open access, voluntary open access, and commons-based approaches. It then assesses the outcomes and policy implications of these approaches. Finally, the chapter concludes by highlighting the need for context-specific considerations of open access applications, particularly when the use of these strategies in the Global South could disrupt existing systems without offering better access conditions or enhancing developmental outcomes.

Despite its common usage in infrastructure regulation, there is no single, formal definition of open access in telecommunications policy and regulation. However, there is some consensus around open access principles in policy and regulation.¹ Common tenets include nondiscrimination and price transparency, wherein nondiscrimination is defined as equal access to networks and wholesale services;² and price transparency, meaning that prices are displayed to the public and customers understand how prices, based on actual costs, have been set.

The rationale for adopting open access regulation in telecommunications is to enhance overall consumer welfare, through both access and price, by improving both market competition and the efficiency of the telecommunications network. In the policy context of universal service, open access theoretically overcomes the problems associated with the high sunk costs required in infrastructure industries that make the duplication of certain network elements uneconomical.

Open access has long been a favored instrument of regulators to guarantee seamless access for consumers to competing networks and services, which nowadays include the Internet. The core principle of open access regulation is nondiscriminatory access to communication-enabling efficiencies, which would create surpluses for investment in network extension. Krämer and Schnurr (2014, 5), on the basis of this broad consensus, define open access in relation to infrastructure as follows:

Open Access regulation refers to the mandated or voluntary provision of access to an upstream resource which must be based on the principle of non-discrimination. The concept may apply to publicly or privately owned access providers that are vertically separated, integrated, or represent a cooperative of multiple entities. Open Access regulation usually refers to the network layer, but may also be applied to other layers of the telecommunications value chain.

A growing body of evidence from more mature regulated markets indicates that the adoption of mandatory open access network strategies may come at the expense of investment and innovation (e.g., Bauer and Bohlin 2008). These trade-offs need to be assessed not only by means of the kind of static efficiency and instrumental competition models that are typically used to regulate the telecommunications sector, but also through dynamic efficiency models more suited to the rapidly changing and fluid Internet environment.

Open access regimes exist, then, to address the interrelated problems of (1) inflated prices, (2) reduced consumption of services as a result, and (3) insufficient investment and innovation. The first two contexts can best be understood in terms of static economic effects (i.e., at a given point in time). The dynamic economic effects associated with the third, innovation and investment, are the most difficult to solve with an open access regime (Gillwald, Odufuwa, Rademan, and Esselaar 2016b).³

As we have moved from a regulated telecommunications environment to that of the unregulated Internet, notions of open as regulated and unregulated access have collided. This is illustrated by the contested terrain of net neutrality, which has evolved with changes in technology and markets, as well as the propensity to regulate them over time. Net neutrality means many things to many people, but, at its core, it is about open access and control. An unwavering principle for proponents of the free and open Internet type of net neutrality is demonstrably a fluid concept in regulation, depending on definitions of the Internet as content or carrier.⁴ The net neutrality debate has its roots in telecommunications regulation, particularly in the context of the US common carrier rules and in liberalizing markets to ensure interconnection with incumbent or dominant players. It also emerged with the popularization of the Internet and the shift in focus to informational resources.

In the net neutrality discourse, the argument has been around retaining the end-to-end architecture that has characterized the free and open Internet and is regarded as neutral; this arguably places limitations on the quality of services and product innovation. The argument put forward by operators and service providers, on the other hand, is for a shift to intelligent network design capable of allocating access to the infrastructure based on the identity of users. The former favors realizing the social value of the Internet, while the latter privileges private value (Goldsmith and Wu 2006).

Most recently, net neutrality has found expression in the zero-rating debate, and the calls by advocacy groups for zero-rating to be banned for fostering discrimination among providers of online content and content applications in ways that may skew incentives for subscribers (i.e., users may choose to access the free services of identified partners instead of the services of competitors). As discussed further later in this chapter, the issue of net neutrality has arisen starkly in the Global South over the zero-rating of Internet services, particularly the offer of free access to social networking platforms (e.g., Facebook’s Free Basics, formerly Internet.org). Net neutrality proponents argue that such zero-rated services create walled gardens that limit access to and use of the free and open Internet. Many advocates have called for Free Basics to be regulated, and even banned. Some have been successful, the highest profile case being that of India.

Yet, from a pricing point of view, the concept of zero-rating has underpinned the expansion of the free and open Internet; in regulation, it has referred to the relationship between the network providers or Internet service providers (ISPs), depending on who is providing the service to the customers (and in mobile broadband markets, such roles are increasingly conflated), and content providers, who are prevented from charging fees for access to their customer base. As pointed out by Lee and Wu (2009, 63), since its inception in academia in the 1980s to its mass popularity in the 1990s and beyond, the Internet has maintained a pricing structure that is unique among information networks: “users and content providers typically pay ISPs access fees—fixed fees to get on the Internet at all—and usage fees—variable fees paid based on time or bandwidth usage; however, there have not generally been any additional charges for one user of the network to reach another user or content provider, reflected in the concept of the peering (free exchange) rather than paid interconnection, associated with traffic exchange in telecommunications networks.”

As the balance of profitability shifted from the network providers and ISPs to the platforms and over-the-top⁵ (OTT) applications (e.g., Skype and WhatsApp), especially ones that replaced high-cost local and international voice calls, telecommunications operators began calling for regulatory intervention that would enable them to charge certain users of the Internet, such as, notably, large application or content providers, additional fees to reach their subscribers, and enable forms of revenue sharing. Research by the European Telecommunications Network Operators’ Association (ETNO 2012), for example, discusses ETNO’s efforts to have regulations on revenue-sharing with OTT services adopted at the World Conference on International Telecommunications.

As Brett Frischmann (2005) recognizes, however, things are not as straightforward as they might appear. Prevention of quality of service discriminates in favor of various applications in the end-to-end architecture debate on net neutrality grounds. Additionally, by extension, it safeguards the commons, which favors data applications, at the expense of time-sensitive applications such as voice and video. The latter are being pushed generally by private and, particularly, giant companies requiring intelligent networks to manage network profitably and are viewed as being not net neutral in these generally polarized debates. But such arguments treat the generation of private value and social value as mutually exclusive. This chapter demonstrates the importance of creating incentives for private investments to deliver social value in countries where public resources do not exist.

Frischmann (2005) further argues that shifting to a quality of service regime where use is determined solely by private property owners that are able to offer such services (including arguably significant social value in the ability to offer greater cybersecurity) also is clearly not net neutral. He highlights the prioritization of innovation in the arguments that are made for optimal Internet design. Private property owners, on the one hand, argue strongly for incentives for innovation in private allocative models that are responsive to market demands. The counterarguments to this are made by those who advocate for the extension of the commons precisely to enable innovation (Lessig 2001). Frischmann (2005, 1008–1016) points out that innovation and competition in upstream and downstream markets, where the debate invariably lands in liberalized markets, is an important part of the policy debate. He notes, however, that “the Internet supports a substantially wider range of socially valuable downstream activities that are neither innovative nor commercial,” have public or nonmarket value, and require equal policy consideration. He concludes, after a lengthy analysis of the arguments of Wu (the main proponent of net neutrality to safeguard a free and open Internet) and Lessig (the main proponent of the innovation commons), that while a theoretically neutral system has significant benefits, true neutrality is not attainable.

Besides the questionable attainability of net neutrality, the debate around neutrality and zero-rating is generally based on assumptions of universal access, or at least widespread access and quality of service thresholds that do not even pertain to the developing countries where such regulatory principles are now being instrumentally adopted.

The zero-rated approach to access contrasts with the typical private provision of the Internet at regulated or unregulated data tariffs. However, free services are not new. The major social platforms, including Facebook and Twitter, are free to users, with a business model based on two-sided markets where marketers and advertisers pay high prices to exploit user data (Bauer 2014; Economides and Tåg 2012). There are few, if any, open goods that are free of cost or nonproprietary in all stages of provision. Thus, as Spence and Smith point out (see chapter 7 of this volume), while free services offered by private companies may be surprising, private companies have long developed innovative ways to use open practices to generate revenue streams and to capture value in other areas of the business, or, in the sum of it, by deploying such strategies.

While the rationale for the monopoly provisioning of communications infrastructure in the form of public utilities was based on demand-side value associated with public goods, as markets became liberalized and the delivery of public services shifted to the private sector, the focus of regulation shifted to the supply-side value associated with private goods, and the extraction of rents. Yet, these were not ordinary private goods; they were public goods delivered by the private sector, and thus they required regulation to ensure that they were delivered affordably (Hirshmann 1958). In both the analysis of these problems and the proposed solutions, “complex motivational structures and … diverse private-for-profit, governmental, and community institutional arrangements that operate at multiple scales to generate productive and innovative, as well as destructive and perverse outcomes (North 1990 and 2005)” (Ostrom 2009, 408), need to be taken into account.

Historically, debates about resource management have been polarized into market versus state control of resources. Markets are narrowly seen as closed or proprietorial, while state ownership is potentially more public and open, although if monopolies are the providers, this is not necessarily the case. Advocates of open access frequently call for the creation or protection of the commons as an alternative to state control. This is, generally, a rejection of the management of resources, even if regulated, through market relations that are associated with private control. However, as pointed out by Frischmann (2005, 1), this “prescriptive call” arising from the revisiting of the commons is underdeveloped both from an economic perspective and in terms of the level required for policy implementation; meanwhile, on the private control side, there is a robust economic theory supporting the market mechanism with minimal government regulation that informs the approach to best practice infrastructure reform, and there is little theoretical support for the implementation of infrastructure commons (Frischmann 2005, 2–5).

Drawing on Ostrom’s theory of common pool resources (Ostrom 2009), Frischmann (2005, 2012) presents an economics-based case for why some classes of key resources need to be managed in an accessible manner. Conventional economic analyses of infrastructure focus primarily on the supply-side value of infrastructure and the profit imperative in network investment and regulation, but Frischmann (2012) explores demand-side considerations to analyze how infrastructure resources generate value for consumers.

Three key insights emerge from this demand-side, value creation–focused analysis. The first, as also found in other infrastructure theories, is that infrastructure resources are fundamental and generate value when used as inputs into a wide range of productive processes. The second highlights that the outputs of infrastructure industries are generally public and nonmarket goods that create positive multipliers in both the economy and society. The third is that “managing infrastructure resources in an openly accessible manner may be socially desirable when it facilitates these downstream activities” (Frischmann 2005, 918).

Frischmann (2005), building on traditional economic concepts used in welfare analyses of infrastructure resources and the societal demand for such resources, puts forward a new theory of infrastructure. He notes that despite the extensive role of private and commercial delivery of information infrastructure, with the increased positive externalities derived from information infrastructure, the role of the state as coordinator and regulator in ensuring its provision and management is still required to ensure widespread access by citizens and the overall distribution of social gains (Frischmann 2005, 919–921).

The following section provides an empirical examination of different open access models, including mandatory-, voluntary-, and commons-based approaches to access regulation of competitive networks, open access broadband networks, free public Wi-Fi, and zero-rated services. The types of open access presented here, and shown in figure 8.1, are all intended to enable or enhance Internet access in developing-country contexts. The analysis of the cases and the conclusions presented thereafter highlight both the need to find new ways of addressing digital inequality and the dangers of static and instrumental regulation.

The first case looks at the application of open access to the mandatory regulation of more traditional telecommunications networks within the context of supply-side valuation associated with economic and competition regulation. Because markets were liberalized and competitors to public monopolies entered the market, open access regulation was mandatorily applied to ensure access to the remaining natural monopoly elements of backbone networks and fixed access network infrastructure. Open access regulation helps to ensure the seamless connection of customers on different networks by mandating the nondiscriminatory interconnection of networks. Facilities are thus available to connect other network operators, especially those that might otherwise have formed bottlenecks, including the wholesale segments of the market (backhaul⁶ and the access network) where few suppliers exist and markets naturally tend toward oligopoly or monopoly.⁷ This has led to the creation of open access regimes that compel dominant operators to provide access to rival operators to enhance competition.

Europe’s open access regulation introduced competition into various monopoly elements of the market by mandating access at multiple layers of the network. This can be implemented through the separation and opening up of network elements, particularly the last mile or access networks, which have been the most uneconomical to duplicate from a fixed-line perspective. Unbundling of the local loop⁸ has been a requirement of European and American regulators to enable competitive access to the last mile. At the backbone level, fair access to networks built with public investments as public utilities prior to any privatization has been accomplished through the structural separation of the network and the services, which are then offered independently to customers.

Local Loop Unbundling One way of implementing open access on traditional copper access networks has been through local loop unbundling (LLU). The rationale here is to foster competition and reduce telecommunications costs by eliminating the duplication of large investments required for last-mile infrastructure. South Africa, as one of the only markets in sub-Saharan Africa with any significant copper networks, has long had LLU on the reform agenda. This is in line with claims made by dominant epistemic communities within the Organisation for Economic Co-operation and Development (OECD), International Telecommunications Union (ITU), and World Bank a decade ago to the effect that LLU was a best practice scenario. The stated argument was that it would enable telecommunications providers to innovate and differentiate their product offerings, promote competition in the provision of broadband services, offer opportunities for innovation to drive product and price differentiation, speed up national economic growth, and even increase competitiveness in the global market and create employment opportunities. With the identification of a second network operator by the South African government after none had met the public policy–determined criteria to be licensed by the regulator, an environment more conducive to new entrants had to be established.⁹ Prevarication on the issue continued for years as a result of the conflict of interest of the government in South Africa. This conflict resulted because the government was responsible for market reform of a sector in which the state held a significant share of the incumbent operator, Telkom (Gillwald 2005), thereby preventing the second network operator from benefiting from market reform of the previous monopoly market.

The public process proceeded at a snail’s pace. This was because the regulator, the Independent Communications Authority of South Africa (ICASA), was facing its own internal leadership crisis. It was also pursuing other urgent activities with more powerful interests (e.g., spectrum) than the second network operator. Despite the publication by the regulator of a findings document identifying four options for unbundling, including an easily implementable bitstream option, ICASA never published any regulations from that round of hearings. This was widely perceived as being a result of the obstructionism by the incumbent, Telkom (Ellipsis 2011).

As all elements of mobile networks are considered replicable (other than termination, which is inherently monopolistic), open access to mobile networks through an enabling application programming interface (API) or API access has not yet been considered mandatory. However, given the dominance of mobile operators in developing-country markets, this is being reviewed in the context of mandatory infrastructure sharing.¹⁰

Functional and Structural Separation Market restructuring through functional and even structural separation is a more drastic form of intervention to deal with incumbency or dominance in the market. Essentially, this restructures the market by separating out vertically integrated operations and compelling them to treat their downstream operations in the same way that they treat their competitors. With the network and services separated structurally, the network is often transformed into an open access, carrier of carriers network that can offer services only under the separated sister company. In South Africa, Telkom, which had been partially privatized in 1996, was required by the Competition Commission to functionally separate its activities as part of the remedies imposed after it was found guilty of anticompetitive practices in 2013. This separation did not restructure the market, but it did give effect to the policy and regulatory framework that required the incumbent to operate its wholesale and retail activities independently enough from each other to allow operators and service providers to compete fairly. Following several proposed turnaround strategies, including complete privatization and renationalization, and in the face of growing competition and price undercutting from new fiber companies, Telkom voluntarily undertook to separate structurally under the mantle of the national broadband champion (Padayachee interview, 2016). Given the voluntary nature of this restructuring exercise, the effects on the market may not be as dramatic as mandatory separation; but the effect on bandwidth prices in the market has been positive, and Telkom’s own balance sheet has improved consistently throughout the early days of this transition despite the harsh economic climate in South Africa (Telkom 2016, 2017).

Outcomes of various forms of mandatory open access regulation are mixed and are context specific, and comparative data is limited. However, there is evidence that while such initiatives might have resulted in short-term positive consumer gains, they can negatively affect investment and innovation (Bauer and Bohlin 2008). In most African jurisdictions with barely any fixed local loops to unbundle, the focus has been instead on extending wireless networks. Since it is more economically feasible to invest in competitive wireless access networks, there has been little effort to require dominant players to provide access to competitors, despite the fact that outside the main urban centers, there is often only one operator network, or possibly two, available to consumers, with very little choice. As yet, very limited attention has been paid to mandating API access to mobile networks, which would have a similar effect to unbundling the fixed local loop. Rather, the focus of policymakers and regulators has been on mandating national roaming by new entrants on incumbent mobile networks. Without very extensive fiber networks in most African countries, open access, state-owned broadband networks have been set up in several countries (e.g., Botswana, South Africa, Tanzania, and Uganda).

National Open Access Broadband Networks (Carriers of Carriers) Competition at the infrastructure level is not viable in many developing economies (indeed, it is not in many developed ones as well). Also, given the high cost of rolling out broadband, states have instead opted to deploy a single national fiber backbone open to all service providers on a nondiscriminatory price basis as a way of enhancing network expansion and access. Australia is one country that has led the way in this regard. Consolidating use on a single backbone enables economies of scale and cost reductions that translate into lower access prices for service providers. Competition then takes place at the service rather than at the network level. This could still foster the efficiencies usually associated with competition and contribute to consumer welfare through bypassing the high input costs for service providers arising from the duplication of facilities or networks.¹¹ State-owned broadband companies have had anticompetitive effects by either squeezing out private-sector investment (e.g., in Tanzania), or by governments requiring the public sector (often the single biggest collective user in many countries) to exclusively use the state-owned broadband services (e.g., in Uganda) (Lumu 2015).

Mandatory open access broadband networks, especially in developing countries where there is already significant private-sector investment, have not been successful. The high sunk cost associated with the rollout of fixed networks is difficult to replicate profitably. Without conscientious efforts to enable competition, the ability to challenge the market dominance of incumbents is limited. The case of South Africa illustrates these challenges.

The fiber-optic networks of the public rail network (Transnet) and the power utility (Eskom) were set aside for the creation of a second network operator at the end of Telkom’s exclusivity in 2003. However, as the second operator was about to be licensed to Neotel in 2006, and broadband was starting to take off, the Department of Public Enterprises withdrew the mandatory setting-aside of holdings so that the transport communication network could create a new, state-owned, open access broadband company, Broadband Infraco (BBI). Since this decision happened without public consultation and cut across the emerging competitive convergence policy, the process was delayed.

Licensed for the provision of wholesale services, BBI became operational in 2009 (see Gillwald 2007 for a detailed case) following additional delays related to the determination of the type of services that the entity would be entitled to offer. BBI has since invested in national and international backbone communications networks and is one of the main investors in the West Africa Cable System (WACS).¹² However, state coordination has been weak and arguably lacking sufficient investment in BBI to deploy even the roughly 12,000 kilometers of backbone planned (not a great distance for a national backbone for a country the size of South Africa). The decision to create an open access network failed to consider the market dominance of the incumbent, which now has over 50,000 kilometers of unduplicated fiber and over 75,000 kilometers in all. Furthermore, international bandwidth capacity in the country had already increased significantly in terms of quality and pricing by then (Gillwald 2007; Nkhereanye interview, 2016).

Initially, there was some concern that BBI would squeeze out private-sector investment. However, having fallen victim to the lack of coordination between the Department of Public Enterprises and the Department of Communications, which was responsible for allocating BBI’s license, BBI could no longer represent the interest of its target clients, the network operators, that in 2009 went on to cobuild an alternative national infrastructure network, while in the process undermining the viability of the BBI business model.¹³

Nigeria announced a similar open access model in 2012, but the process has been delayed. So far, there has not been an intervention in wholesale markets, and market regulation based on a comprehensive study is long overdue (see the Nigeria case in Gillwald et al. 2016a).

National Open Access Wireless Networks Within infrastructure industry policy and regulation, the creation or enforcement of open access wireless networks has been identified as a mechanism to reduce infrastructure that is not economical to duplicate, and also as a remedy for monopoly. Open access networks have been tried in Kenya and Mexico to counter the extreme dominance of incumbents (above 80 percent of market share), and where such primary markets for achieving universal service were deemed uncompetitive. In these cases, the implementation of open access often takes the form of a consortium-owned mobile network. Here, spectrum is to be shared by any operator or service provider who wishes to offer wireless services on the designated spectrum. In Kenya, the dominant network operator Safaricom backed out of the deal before the proposed network became operational resulting in the collapse of the consortium. In the case of Mexico (discussed next), only the digital dividend spectrum has been set aside for this experiment (Gillwald et al. 2016a).

In Mexico, which is the only remaining country pursuing open access wireless, there is considerable concern about the risks emerging from the constitutional obligations to continue with the process in the face of growing evidence that it might not ultimately fulfill its mandate. A state company, Red Compartida, has been established to hold the spectrum licenses in an attempt to mitigate the risk of failure (Borjón Figueroa interview, 2016). In the meantime, other regulatory reforms in Mexico, which enabled the introduction of classical asymmetrical remedies for the market in which operators were exercising significant market power and distorting market outcomes, were implemented during the six years that it took to finally set up the open access wireless consortium (Labardini interview, 2016), reducing the dominance of the dominant operators to around 60 percent (from 80 percent).

South Africa’s Integrated ICT White Paper (RSA 2016) proposes the introduction of an exclusive wireless network, not limited to the digital dividend spectrum, as proposed by Mexico and Kenya, but rather extending across the entire high-demand spectrum that is currently unassigned (2.6 GHz and the 700 MHz digital dividend spectrum). It also proposes the incorporation of other mobile spectrums when the operators’ 15-year licenses expire so that the full spectrum is held in the open access wireless network (RSA 2016). As discussed previously, the extreme complexities of such proposals require a sophisticated understanding of the mobile market.

Although the white paper provides no insight into the business plan or ownership structure of the wireless open access network (WOAN), a consortium of the kind embarked on for the successful, voluntary, and partially open access undersea cable company consortia (WACS, Seacom, and Eassy) has been suggested as a model. The business model for the construction and maintenance of a single end-to-end cable with a 25-year life span, however, cannot be readily applied to a common carrier model in the dynamic and fast-changing mobile wireless market, where operators’ control of the spectrum determines the competitive advantage.

International expert inputs into the consultative process for SA Connect, the national broadband plan, which was unpublished but publicly available at the time, contended that at least one of the bigger mobile operators needs to be inside the network for it to enjoy economies of scale and scope. Consultants appointed by the National Treasury of the Republic of South Africa to establish the viability of the WOAN also cautioned against its establishment (Bedi and Schumann 2013). To work, the model requires incentives to make the investment in the open access network more attractive to the mobile operators than investing and controlling their own networks. In another scenario, if operators are unable to access any of the additional long-term evolution (LTE) spectrum necessary for the cost-effective development of their businesses (as has been the case in South Africa), then there is no need for them to invest in their own networks. Instead, they can simply access wholesale networks at regulated open access prices.

There are at least two problems: first, capital for the open access network has to come from somewhere, fast; and second, the history of the public open access broadband company does not bode well. With some pressure from government, but also given the promise of low-cost access to BBI (the new state-owned company), the dominant mobile operators, MTN Group Limited (MTN) and Vodacom, opted not to build out fiber networks once the incumbent’s exclusivity period expired, and they were permitted to do so. However, the delays to the licensing process, and subsequently the operationalization of the license and the clear undercapitalization of it for national champion purposes, drove MTN, Vodacom, and the second network operator, Neotel, to build their own intercity transmission network. Because the government again is trying to cut across the policy framework (flawed as it may be, it nevertheless is the result of a consultative process to force an exclusive wireless network), it appears that no lessons have been learned from its own recent history.

The proposed open access wholesale wireless network in South Africa could result in negative unintended consequences, particularly the inhibition of the current extremely high level of private investment in mobile network extension and upgrades. The opportunity costs of this to the country, already lagging on 4G¹⁴ deployment, are potentially enormous and include negative impacts on economic growth and employment, which would take years to correct.

State efforts to intervene in the successful mobile market with mandatory open access wireless networks for the high-demand spectrum in South Africa may be counterproductive if the goal is to meet development objectives. Furthermore, there is anxiety about whether there is the institutional capacity and sophistication to oversee the complexity of creating a viable, single, open access network. After all, the absence of state coordination in relation to the spectrum has plagued far simpler processes. The WOAN may well fail for the same reasons that complex auctions, as an alternative to open access and other models, may also fail. The lack of success of any auction in Africa is cited by open access advocates (Song interview, 2017; Rey-Moreno interview, 2017) to argue against the multilateral best practice orthodoxy on spectrum assignment of auctions.

There are a couple of rejoinders to this position. There are few ways, other than auctions, of valuing and assigning spectrum set aside for commercial use competitively and in ways that will ensure optimal efficiency in the use of the spectrum (including regulated spectrum trading to enable the correction of any auction outcome errors). For example, the South African Treasury is regarded as a highly effective government department, and the high standing of the financial sector in South Africa and the long view taken on the economy is attributed to them. The Treasury already provides technical assistance to most departments through its Government Technical Assistance Committee (GTAC). Within policy and regulatory confines that would contain fiscal interests in maximizing the pricing of the spectrum, the department could assist the regulator with the technical dimensions of the auction and the management of any external consultants who are brought in. This could be implemented far more quickly than an exclusive open access network, which is bound to face multiple rounds of legal review before the network could even be established and become operational. The release of the high-demand spectrum has already been delayed by six years.

The best case scenario if legal hoops were navigated, using the Mexican model as some measure, even with some lessons learned, is probably a minimum of five years. That would already place South Africa ten years behind early adopters of LTE in the allocated and cost-effective bands (although operators have refarmed alternative, less-cost-effective spectrum that they hold currently to make it available for LTE). All operators are offering such services in the major centers. With the urgent release of high-demand spectrum, and with all the previously mentioned caveats for making available unlicensed and open access spectrum, regulators should be focusing on creating incentives for infrastructure sharing, channeling complementary investments, and moving to an industry-coordinating role by enabling voluntary models that allow the participants to assume the risk associated with high sunk-cost infrastructure investments.

Voluntary Open Access International Backbone Networks (Commercial Models) An example of successful voluntary open access in Africa is the provisioning of undersea international bandwidth by Seacom (and arguably WACS as well). Seacom broke the monopoly that SAT-3 (South Atlantic 3 submarine cable) had on Africa operating as a closed club consortium consisting of incumbents. The commercial logic of Seacom, and to some degree the shared infrastructure consortium that operates WACS, is similar to that of the commercial, open access, dark fiber national transmission networks (described next), and some have argued that that should be the form of the wireless open access network in South Africa.

Voluntary Open Access Fiber Networks (Commercial) With the opening of backbone competition, commercially operated, voluntary, open access networks operate in the South African market in complementary ways with existing broadband network investments and OTT platforms and applications. Commercial companies that operate voluntary open access models have managed to optimize their traffic flows and maximize their return on investments in as short a period as possible, allowing them to refinance the next phase of network rollout and contribute to the cost of network extension.

Open access companies such as Dark Fibre Africa (DFA) and FibreCo created competitive alternatives to Telkom on main intercity routes (MyBroadband 2014). This competitive pressure in the national data transmission market compelled Telkom to review its strategy in the market and adopt an open access model, undergoing a voluntary structural separation of its wholesale and retail divisions, as already mentioned. This has largely been welcomed, although it has not entirely eradicated the lingering skepticism of the incumbent’s competitors and clients represented by the industry association that brought the initial anticompetitive case against Telkom. A voluntary regime could be reversed at any time, and there is still a view among some stakeholders that, given that Telkom is the dominant backhaul provider, oversight of the open access regime by the regulator is still necessary (Cull interview, 2016; Cohen interview, 2016; Nkereanye interview, 2016).

This open competitive environment results in the duplication of networks, certainly in the metro areas where competition can be easily sustained. However, there is also extensive evidence of complementary investments by different tiers and types of operators, and procurement or sharing of infrastructure for primary and redundancy use for national coverage (Hussein interview, 2016).

As the acting director general of the then-Department of Telecommunications and Postal Services (DTPS) in South Africa¹⁵ indicated, commercial fiber “has been one of the most phenomenal developments in the sector, a game changer that demonstrated that open access networks are viable, unlike what the traditional operators have argued” (Mjwara interview, 2016).

The success of voluntary, open access, commercial fiber networks demonstrates that delivery of social goods through supply-side value can be enabled through demand-side valuation of those services and the creation of incentives for operators, such as the government anchor tenancy identified in the South African national broadband policy, SA Connect. It was the success of this voluntary open access model that led SA Connect to propose that instead of building another state-owned network for the public sector, as was being suggested in government circles but for which there was no state funding, that the extensive private networks be provided with the incentives to extend their nearby networks to underserved areas. The proposal was to aggregate public-sector demand and smart procure services from existing providers at cost-effective prices. In this way, incentives such as government anchor tenancies through demand aggregation in underserved areas could foster broadband investment in areas that would otherwise be uneconomical to service.

Despite this potential of leveraging private investment to furnish public-sector broadband services as proposed in SA Connect, little to no progress has been made since 2013 in this regard. The lack of capacity within the ministry to coordinate the plan across government departments is the main stumbling block.

Free Public Wi-Fi—Bring Your Own Device The use of universal service funds to reach underserved communities and to provide access to the Internet was locked into fixed-lined models for a long time, and today it remains largely focused on subsidizing network extension and the aggregation of demand through computer-based telecenters. However, with the advent of mobile broadband and smart devices, the price and skill barriers that the computer-based Internet access created were essentially removed. This undermined the logic of access aggregation of this kind. Mobile devices can now offer adequate user experiences, with web browsing, email access, messaging, and an increasing variety of applications for using the Internet without having to travel to single points of presence.

While increased Internet penetration (with its widening coverage), mobile technology upgrades to 3G (and now LTE), and the relative affordability of smart devices have combined to increase mobile data uptake, the majority of people in the Global South still lack access (see www.afteraccess.net). Limited coverage and slow data speeds outside the major centers, together with the relatively high cost of services, exclude the poor and inhibit the optimal use of mobile data by most users (Rey-Moreno 2013; Stork, Calandro, and Gillwald 2013). Yet in only a few countries have universal service funds or agencies made this transition to supporting user access.

Nevertheless, in some countries, such as Rwanda and South Africa, metropolitan governments are offering free public Wi-Fi (FPW) as part of their strategy to provide citizens with connectivity to access government services and the Internet more widely. Given the widespread use of Internet-enabled phones with sufficient computer power and adequate screen sizes for meaningful use (even among lower-income groups), FPW is a promising solution. FPW sponsored by a local authority is popular in many major Asian and North American cities; and the South African broadband plan, SA Connect, requires that all public buildings be connected to broadband and that, whenever these connections occur, they include a FPW hot spot. But provinces and cities in South Africa are moving even more swiftly by actively investing in both central business districts and underserviced areas, rolling out FPW. The traction with citizens has been overwhelming.

There are various models of public Wi-Fi in South Africa, but the two largest initiatives, in Gauteng and the Western Cape provinces, respectively, implemented assessable initiatives by 2016. These are the wealthiest provinces in the country and are highly urbanized (Gauteng almost entirely). There are projects in the four largest metropolitan areas: Tshwane (formerly Pretoria); Johannesburg and Ekurhuleni (in Gauteng); and Cape Town (in the Western Cape).

FPW projects typically emanate from authorities that already have broadband development strategies and plans in place, as is the case in both Gauteng and the Western Cape. Most FPW plans are linked to initiatives to interconnect government buildings via fiber networks. FPW projects tend either to piggyback on these fiber deployments (Western Cape, Ekurhuleni, Johannesburg, and, to some extent, eThekwini in KwaZulu-Natal province), or to benefit from existing municipal fiber (Tshwane and Cape Town) (Geerdts et al. 2016).

Of all the projects, the state-funded Tshwane Free Wi-Fi, Project Isizwe, is the most advanced in terms of scale and impact. The project began in 2013 and is based on two investment justifications: first, broadband, as a basic right for every citizen, should be offered across the country, fully funded by government; and, second, increased broadband penetration will drive economic growth, increase commercial and financial activity that will expand the tax base, and altogether will exceed the city’s initial investment. On this basis, public Wi-Fi should be furnished for free (Geerdts and Gill-wald 2017).

The results, in terms of network deployment, have been impressive. The content portal and program launched to generate local video illustrate the potential for Wi-Fi and broadband to enhance employment opportunities, improve health and education, and connect government to citizens. In line with the national broadband plan, SA Connect, the focus has been on connecting educational institutions.

In the longer term, Project Isizwe expects that input pricing will reduce to the point where free Wi-Fi is no longer considered a significant expense. Until then, the model is entirely dependent on substantial public funding, which has been large for a relatively small-scale, low-cost, public Wi-Fi network, at around $75,000 a month (Geerdts and Gillwald 2017).

The model for the Western Cape is more complex. The province and city take a holistic view of promoting broadband on multiple fronts. One of the foundations of this model is to stimulate private-sector investment in broadband by improving market information (RIA 2016) to help investors understand the costs, by creating demand (with government as an anchor tenant), and by reducing the capital outlay requirements (by paying for part of the infrastructure). At the provincial level, the government is implementing a fiber/Wi-Fi contract to connect government buildings and selected Wi-Fi sites. What was intended to be a public-private open access network was scuttled by the national State IT Agency (SITA), which asserted its prerogative to undertake a regular competitive tender under its auspices. As a result, a privately operated, closed network was created, with premiums significantly above those offered by the public open access network initially proposed.¹⁶ The winning proposal allowed the second network operator, Neotel, to deliver services to government, as the anchor tenant, via both fiber and Wi-Fi, but also to exploit the private-sector market (Geerdts et al. 2016).

On the other hand, the City of Cape Town (CoCT) is encouraging private investment by reducing the upfront capital outlay required. The CoCT provides connectivity from a central point to remote locations while also supporting access points (APs). The ISP only has to offer data connectivity. This model is more expensive to implement than the Tshwane model, but includes the private sector and fosters innovation and competition. See comparisons in table 8.1.

What this natural experiment demonstrates is that various models of public Wi-Fi can be justified on the basis of the demand-side value of the network, as the Tshwane municipality has done to justify the substantial subsidy that it provides to offer limited free access to the Internet. Although the supply-side value of public Wi-Fi has been more difficult to institute across the world, the Western Cape government has attempted to support the private provisioning of public goods by enabling access to low-cost bandwidth and providing start-up subsidies, but, ultimately, it intends them to become self-sustaining and economically viable. See comparisons in table 8.1.

Free, Limited Access and Use: Zero-Rating Services Zero-rating has been defined as tariff plans that enable mobile (wireless) customers to download and upload online content without incurring data usage charges or having their usage counted against data usage limits (Eisenach 2015). In developing countries, zero-rating allows mobile subscribers to access certain or preselected online content for free (i.e., without having the associated data usage counted against their usage allowances under wireless service plans). Zero-rating practices, therefore, are considered a violation of the net neutrality principle by some regulators. In India, this led to the mobilization of civil society to produce a million-signature campaign, which subsequently led to the banning of Free Basics and other zero-rated products by the Indian regulator, Telecom Regulatory Authority of India (TRAI). Most African countries, which tend to follow European Union (EU) regulatory trends rather than those in the United States, have been slow to develop positions on net neutrality, with the EU itself only adopting net neutrality rules late in 2015. The EU rules enshrine the principle of net neutrality in law: no blocking or throttling of online content, applications, and services. Under these rules, efforts to prohibit ISPs from blocking, throttling, and discrimination with regard to Internet traffic are largely focused on the quality of the service and treatment of traffic.

Services that are typically zero-rated by providers in developing countries include the world’s biggest web properties, such as Google, Facebook, and Twitter, as well as messaging services such as WhatsApp, KakaoTalk, and WeChat that can reduce the high cost of communicating through phone calls and short message service (SMS)/text messages. See the comparisons in table 8.2.

The empirical assessment of the cases demonstrates how the long history of mandatory open access regulation, derived from both regulation in the EU and multilateral agency reform programs as a way to enable market entry and manage market dominance and anticompetitive practices, has not transformed easily into fixed broadband network extension in South Africa and Nigeria, an open access wireless network extension in Kenya, or, as yet, South Africa (e.g., Gillwald et al. 2017).

All the national open access broadband cases highlight the institutional constraints on implementation, with the failure to implement since 2012, the planned wholesale open access national broadband network as cornerstones of national public policy. Inadequate funding and insufficient institutional and regulatory capacity to create a level playing field for the new entrants and investors, especially in upstream segments, explain this failure. See the comparisons in table 8.3.

State-Owned Open Access Wholesale Broadband Companies The introduction of open access wholesale networks in countries with very low per-capita incomes or small population sizes, such as Tanzania and Botswana, in countries where the duplication of competitive networks may not be feasible, and in countries unable to attract competitive private investments may have resulted in the extension of broadband networks across the country. As shown, however, this does not necessarily mean that such networks meet national objectives of digital inclusion, either because of the unavailability of access networks or consumers’ inability to afford the Internet-enabled devices needed to access them.

The cases of state-initiated, open access, wholesale networks examined here demonstrate that the more sophisticated mandatory open access arrangements requiring institutional expertise to control or manage resources are limited or entirely absent in many developing countries. Even in more dynamic markets, such as Kenya, Nigeria, and South Africa, where state-owned entities needed to compete with or attract private operators, they have largely failed. In South Africa, the national broadband company never really thrived due to undercapitalization and inability to compete in the dynamic and fast-growing voluntary open access market. In Kenya and Nigeria, state-owned wireless and regional wholesale broadband networks proposed in those countries’ national development plans were never even implemented.

Early efforts to intervene in the successful mobile market with early mandatory open access wireless networks for high demand spectrum in Kenya and South Africa have faced similar institutional constraints, perhaps even more so than in the case of fixed broadband, where there was not much competition at the time of the initial interventions. Open access initiatives in the extremely successful mobile sector, where billions of dollars are being invested in broadband network extension while access prices are coming down, may turn out to be counterproductive in terms of meeting development objectives.

The voluntary open access undersea cable company consortia (WACS, Seacom, and Eassy) successfully broke the dominance of the club consortia of incumbent operators that monopolized the provisioning of international bandwidth, the high cost of which has prevented Internet proliferation in Africa for decades. A similar consortium model applied in the dynamic mobile market will require clear incentives to make the investment in open access networks more attractive than operators investing in and controlling their own networks. In all the countries assessed in the case studies, the dominant mobile operators are also the largest investors in broadband rollout. If they are permitted to access the high demand spectrum while continuing to operate their own networks, then any competitive open access network will most likely be marginalized. On the other hand, if mobile operators are not permitted to access the desired spectrum in an attempt to compel them to use the WOANs, then there is no need for them to invest in the latter, as they can simply access wholesale networks at regulated open access prices instead of investing in a shared pool. This is the impasse faced in South Africa.

Additionally, anxiety persists about the existence of the institutional capacity and sophistication necessary to oversee the complexity of creating a viable single, open access network. The absence of state coordination in relation to broadband, first with the state-owned company, BBI, and then in relation to spectrum, is evidence of the institutional risks associated with the management of complex processes. The decade-long digital television migration debacle in South Africa, which has now been overtaken by global events, together with the six-year ministerial delay in issuing a policy directive on the release of high-demand spectrum, which has now resulted in a legal standoff between the Ministry of Telecommunications and Postal Services and the regulator, provide evidence of both the complexity and risk of this undertaking. The traditional asymmetries of information between operators, regulators, and the ministry that have plagued far simpler processes, such as licensing hearings or the successful completion of market reviews, are unlikely to be overcome in the short run.

In the context of limited capacity and resources, alternative approaches that leverage private capital and skills, lower regulatory risk, and use the large public-sector demand for broadband as an incentive for investment by the private sector are required. The unintended consequences and potential policy and regulatory failure in Mexico, Kenya, and Rwanda, all with far less competitive and far more concentrated markets than South Africa, should be taken as a caution to policymakers. Without the institutional capacity and sophistication to oversee the complexity of creating a viable, single, open access network, more stringent enforcement of competition in the market to deal with dominance through well-established practices of asymmetrical economic regulation may be more feasible, as the case of Mexico demonstrates.

Finally, the high levels of risk associated with the likely protracted legal battles that would ensue from an exclusive open access network, as proposed in South Africa, would delay the release of high-demand spectrum, at considerable cost to the economy.

Voluntary Open Access Fiber Networks The commercial companies in South Africa that operate voluntary open access models to optimize their traffic flows and maximize their short-term return on investments so that they can refinance the next phase of network rollout are contributing to network extension in a cost-effective manner. Although this results in the duplication of networks, this is certainly the case in metropolitan areas, where competition can be easily sustained, and there is also ample evidence of complementary geographic investments by different tiers and types of operators. Moreover, procurement or sharing of infrastructure for primary and redundant use for national coverage is taking place.

Accepting that large numbers of Africans will not be able to afford to be optimally online, even if markets were effectively regulated and prices were cost based, deploying spectrum to create and extend the commons (i.e., the unlicensed spectrum) would be a key enabler. Extending commercially available public Wi-Fi from elite urban areas, possibly through deploying poorly utilized universal service funds or other public resources to all public spaces, offers a viable way of increasing the intensity of use in urban areas and enhancing network effects that would contribute to more inclusive digital development.

This chapter argues for an innovative policymaking approach that understands the need for a new interplay between state and market. It looks at novel models of access, service delivery, investment, and risk that leverage private and community knowhow, low-cost technology innovations, and complementary access solutions such as FPW.

There is a need for even greater regulatory agility and insight to manage tensions between the various policy objectives of competitive efficiency, innovation, and consumer welfare, as well as the safeguarding of the public and social value of the Internet through the extension of spectrum commons (i.e., an unlicensed and social use spectrum). In developed and developing countries alike, most of the spectrum is largely unused outside main metropolitan areas. In the sharing economy of the Internet era, we are already seeing voluntary infrastructure sharing by operators. These types of approaches need to be embraced by governments from a critical resources management perspective. Enabling secondary spectrum use would enable new dynamic spectrum sharing, which operates at a fraction of the cost of the Global System for Mobile Communications (GSM) network (originally known as Groupe Spécial Mobile) to be deployed on new business models in the largely unused spectrum in rural areas. Such an approach could instantly provide low-cost, high-quality bandwidth.

With the long-term evolution of fifth generation (5G) under way, African governments need to ensure that the potential of this technology, which operates within a spectrum-sharing environment with data offloads to proprietorial and open public Wi-Fi, is harnessed for public purposes, not just for niche commercial applications.

Similarly, one of the main arguments in favor of zero-rating is that it brings down the cost of access to and use of information in developing countries. A user of Wikipedia Zero, for example, has unlimited, no-cost access to everything in the online encyclopedia. Furthermore, providing free access to popular content and services is preferable, from an access to information perspective, to restricted access or no access at all. Such free access may drive the demand for general-purpose, mobile Internet access that can stimulate demand and fund investment in infrastructure. If the unverified data provided by Facebook can be used as an indication, it may be that zero-rated services provide, to some extent, a gateway to the Internet. Facebook claims that 50 percent of Internet.org / Free Basics users move on to use some paid data service within a month of using the free service for the first time (Internet.org 2015¹⁷ quoted in Futter and Gillwald 2015).

Traditionally, the net neutrality argument has been applied to ensuring equivalent quality of service to everyone who accesses the Internet by preventing positive pricing discrimination. The main problem with the net neutrality argument for banning zero-rated products, however, is that it prioritizes the principle of net neutrality over other public interest principles of universality and equality (of access, not quality) (Gillwald et al. 2016a). Applying net neutrality to zero-rating through negative pricing discrimination affects not the technical quality of the Internet, but the entry to and use of it. In African countries, where affordable access is the main factor inhibiting Internet take-up, and where even cost based prices may be unaffordable to many, zero-rated services may provide access to the Internet that otherwise would not be at all possible.

The critique of zero-rating on the grounds of breaching net neutrality in the mobile prepaid environment also tends to conflate different competitive outcomes for different players and elements, which, if found to have anticompetitive outcomes, may require different policy and regulatory remedies to address them at the level at which they occur, rather than through blanket bans or restrictions.

RIA’s detailed pricing of products across the African markets also confirms that zero-rating is one of multiple short term market strategies to grow market share by late market entrants. If most users do not buy data at some point, it will not make sense as a business strategy for mobile network operators to continue to discount their data use of this platform (Gillwald et al. 2016a).

With regard to the mobile market, which is the market providing access to the majority of Internet users in Africa, zero-rating of products by late entrant operators in fact may be competition enhancing; and regulators should refrain from regulating it, so long as it does not establish or entrench anticompetitive practices or long-term dominance in a market. While it can be argued that zero-rating enhances the dominance of the platform that is being supported (e.g., Free Basics) and channels users to it, it should be noted that social networking in general, and Facebook in particular (whether it is zero-rated or not), is driving Internet take-up.

The cases of open access discussed in this chapter highlight the need for a case-by-case, context-specific consideration of open access applications. In highly resource-constrained environments, those resources that are available must be optimized to deliver on the social good promised by broadband and digital economy policies.

While open access policies and strategies have again become part of the best practice telecommunications reform templates to deal with absence or bottlenecks in broadband networks,¹⁸ and the principle of openness underpins the public Internet, the uncritical use of open access strategies in developing countries operating under conditions of scarcity and constraint can end up disrupting existing systems without offering more favorable access conditions or enhancing planned developmental outcomes. Network competition is still seen as producing the best outcomes regarding public policy objectives in more developed economies; but the duplication of infrastructure and the fragmentation of demand typical in such cases are not feasible in most developing markets, or even in some more mature ones (Gillwald et al. 2016b). Where states are weak and lack the institutional capacity to enable or enforce open access, the instrumental application of open access solutions may have unintended consequences or promote elite outcomes that can end up exacerbating current bottlenecks or digital exclusion.

In much of the literature on open access and in the advocacy that accompanies its implementation, the value attached to the infrastructure tends to be siloed according to the delivery mechanism. Yet the information infrastructure underpinning our modern economy and society is now a mixed commercial, public, and social infrastructure. In practice, communications resource management has become far more complex than ever, with value being delivered through various hybrid regulated and unregulated models. Communications infrastructure is no longer simply a national communications system, manageable through national policy or regulation. Policy, regulation, and practice develop in response to technology developments in global markets, and increasingly through global governance, although they play out at the national and local levels. We see public and private networks delivering market goods with social value (productive use in downstream activities) and using nonmarket and public goods to transact and produce private goods, as people move from simply being users or producers to being both within the new forms of content development and provision in the platform economy. While the current practices largely undervalue public and social value, as this is classically understood in public goods or the commons, considerable social and public value is offered by private infrastructure investments driven by profit imperatives. This is particularly so in places where, without private provisioning, there was not, or would not be, much delivery of communications infrastructure at all.

Policymakers and regulators need to gear themselves up to the complex adaptive systems that have developed globally; regulators should be focusing on enabling and creating incentives to ensure that the underlying network necessary to support the high-value informational infrastructure is affordably available. This will require reducing cost drivers in developing markets through the regulation of infrastructure sharing through incentives and enabling rights of way regimes, as well as enforcing limitations on fruitless duplications in trenching and masts. Instead of adopting instrumental static regulation of competition derived from telecommunications regulation, regulators need to see how they can be channeling complementary investments and moving toward an industry-coordinating role by enabling voluntary models that allow the participants to assume the risk associated with high-sunk-cost infrastructure investments and open innovation.

The case studies used in this chapter draw on research undertaken for (1) Research ICT Africa policy papers on open access networks, by Alison Gillwald and with Fola Odufuwa, Broc Rademan, and Steve Esselaar; (2) on public Wi-Fi deployment in South Africa, by Alison Gillwald with Christopher Geerdts, Chenai Chair, Enrico Calandro, Mpho Moyo, and Broc Rademan; and (3) on zero-rating, by the author with Chenai Chair, Ariel Futter, Fola Odufuwa, Kweku Korateng, and John Walabengu. For more information, see http://www.researrchictafrica.net.

The author thanks the editors and reviewers of this chapter for their invaluable comments, and especially Raul Zambrano, who contributed extensively to the chapter’s structure. Any errors or omissions are those of the author.

The Internet and other information and communication technologies (ICTs) are transforming the ways in which people communicate and interact. People are now actively participating and interacting online in ways that are unthinkable in the physical world and were unimaginable just a decade ago.

In this context, the concept of open practices has arisen in the literature (see chapter 2 of this volume and Smith and Seward 2017). Open practices are a specific set of ways that people engage and participate online that involve collaborative production processes, as well as the distribution and use of free content. As articulated throughout this volume, under the right conditions, these practices have the potential to help achieve human development targets.

Yet, despite the promised benefits that these practices bring, and the great optimism with which they are sometimes treated in the literature, the reality is that society as a whole is far from benefiting equally from open activities because a large sector of the population is excluded from said benefits (Kularski and Moller 2012; Fairlie 2017; see also chapters 1 and 5 of this volume). As discussed in chapter 2, key features of what is called open are that in theory, there is no direct cost for participating in a certain platform, and anybody can do it. In practice, however, there are barriers such as hidden costs and skills that are needed in order to participate. Participation typically assumes access to the Internet, or at least a mobile network connection, and that users have reached a level of education that allows them to engage in a meaningful manner. In some cases, it is even necessary to belong to a certain social circle to even hear of the possibility of participation. The upshot of these factors is that the benefits of open practices do not accrue equitably.

In this chapter, we explore the relationship of personal factors surrounding the use of ICTs and the extent to which and how someone benefits from open activities or, conversely, remains excluded. The data used in this chapter come from the After Access Survey–2017 carried out by the DIRSI¹ network. The survey collected information about access to and use of the Internet for five Latin American countries, each with a different per-capita income level: Argentina (high-income), Colombia and Peru (upper-middle-income), Paraguay (lower-middle-income), and Guatemala (low-income).

The data and analysis show that the socioeconomic context in which people are embedded, which affects a broad range of issues, from where people live, to education status, to how they access the Internet, to work opportunities, matters a great deal. In particular, we explore how the devices that people use and the places where people access the Internet, coupled with personal characteristics, affect their engagement and potential benefit from open activities.

Two main findings stand out. First, the more people in these countries engage in more open activities, the more familiar they are with the Internet, as reflected by the number of years they have used the Internet, or the more devices they can use. The second result is that socioeconomic context still matters: people with higher levels of education, who have a higher socioeconomic status, or who live in richer countries will engage more in open activities. The first finding gives us room to recommend sector-specific policies. The second finding leaves us recommending sound macroeconomic policies.

The chapter unfolds as follows. The next section presents the theoretical framework in which we discuss the meaning of the term open and outlines the contextual elements that we have identified. The three subsequent sections provide an analysis of the data on the open use practices in the five countries. The first of these defines user profiles based on socioeconomic information and how they access the Internet. The second presents a descriptive analysis of the effect of the diverse Internet access forms on the number of open practices that the agents engage in, associated with educational purposes, government relations, job search, entertainment, and current events. Finally, the third presents our detailed analysis, using a multivariate econometric model, which outlines the impact of context (personal characteristics and type of access) on the probability that individuals engage in open practices, specifically focusing on education, government, and job search. The final section offers our concluding thoughts.

In this section, we define what we mean by open, drawing on the analytical framework proposed in chapter 2. This is a necessary first step in trying to understand how the context (in this case, one’s personal characteristics and type of access) shapes the ways in which one benefits from open practices (or not). After the clarification of the term open, we then explain the way in which the context creates barriers or enables our use of digital platforms, the Internet, and open practices. This gives content to the context, and therefore to the mechanisms that condition the use of technology and the practice of open activities.

This chapter focuses on the open practice of open consumption, understood as the liberty to use content and platforms created and distributed by other people or organizations. By focusing on use, we do not consider how people contribute to the creation or expansion of the platform, but we do emphasize the benefits they receive (and perceive) from open activities. According to Smith and Seward (2017), “Consumption of content is ultimately what allows people to benefit (or not) from open processes, and through those benefits, realize other impacts, such as saving money or achieving better grades.” In this way, the consumption of open activities not only determines who benefits, at least directly, from the open process, but it also shapes other relevant social impacts. Critically, open consumption is highly influenced by the individual’s context, which we define as the means of access and some individual personal characteristics, as shown in figure 9.1.

Before developing a more accurate definition of context, it is necessary to explain what we mean by “benefits of open consumption.” As mentioned, the term open consumption refers to people’s usage of free online platforms or freely distributed resources. Therefore, the actual use of a platform or online resource in a specific area, such as education, job searching, work, or entertainment, implies a benefit. In this chapter, the number of open tasks that a person engages in serves as a proxy of the benefits that one obtains from open consumption. For example, if person A enrolls in a free online course, checks free digital libraries, and accesses a free database, while person B reads literature online, we could argue that person A is receiving more benefits from open education consumption than person B. Of course, the proxy is not perfect; for example, person’s B usage could be more complex and in-depth, while person A’s usage quite superficial.

The Content of the Context: From ICT Access to the Use and

Appropriation of Digital Benefits

In this section, we develop two analytical models that complement the concepts proposed previously. These models explain the role that personal characteristics and the form of access play in the appropriation of the benefits obtained through open activities.

The first analytical model is the one elaborated by Selwyn (2010 and 2015), in which the connection between the use of technologies and the acquisition of relevant variables influences the welfare of people. According to the model, simply having access to a set of technologies does not guarantee the appropriation of the benefits that these technologies can bring. Access is only the first step in the acquisition process.

Dodel (2015) proposed two additional levels—usage and appropriation—that are required before achieving the positive results of technology-based activities. Usage of ICTs concerns the frequency, the familiarity, and the diversity of digital uses that an individual can have. Appropriation means that the individual not only uses the technology, but also understands how the digital system with which he or she interacts works, what benefits can be gained, what damage it can cause them, and what their role is in that interaction structure. The addition of these levels helps illustrate why the consumption of content and use of digital platforms do not have a homogeneous impact on all individuals. The benefits depend on the familiarity they have with the platform, their level of education, their socioeconomic characteristics, and other factors. See figure 9.2 for a summary of this process.

Based on the conditions necessary for the digital benefits appropriation process, Van Dijk and Van Deursen (2014) propose something similar to Dodel (2015) and Selwyn (2010) (see figure 9.3). Van Dijk and Van Deursen suggest that four dimensions are required before achieving the expected benefits: motivation, material access, abilities, and time. The first dimension is the individual’s motivation to use the technologies. For example, some people may have negative perceptions about the use of ICTs or of a particular service, and so they can be reluctant to utilize them. The second dimension is the physical access to the necessary infrastructure, which is still a relevant issue for developing countries. For instance, in the countries analyzed in this chapter, despite the fact that their teledensities exceed 100, there are still gaps in smartphone, tablet, and laptop access, which leaves a significant group of the population at a disadvantage.

The third element refers to the abilities that people have to utilize the devices. The authors mention five groups of relevant activities: strategic (knowing which platform to use for which end), creation of content, communication, operational, and news gathering. People who lack these abilities will not be able to obtain the benefits that the digital platforms can provide. The final dimension refers to the time available for the individuals to use the technologies. Van Dijk and Van Deursen (2014, 2) label this latter process as “the new digital divide,” as opposed to the digital divide, which takes into account only access to ICT devices.

These models assume a neutral access dimension. In other words, they do not differentiate in terms of the kinds of access (a computer versus a feature phone, or at home as opposed to at a cybercafé). However, the way in which users access the digital infrastructure is not homogeneous; rather, it is differentiated by geographic location and/or the different devices users possess or use, as shown in figure 9.4. For example, for one individual to access the Internet from her or his home is not the same as the same individual accessing it from a workplace or school. In the latter, individuals may interact with others and learn about the Internet in ways not available when accessing it alone at home, as explained by Mazimpaka, Mugiraneza and Thioune (2015) and Rabab’ah et al. (2015). This way, certain locations will allow the development of a particular set of skills, while other locations might not. This differentiated access could have a similar effect when considering types of devices: mobile phones may represent a higher level of use of types of certain platforms (mobile applications), while laptops or desktops provide the ability to interact in different ways. As Donner (2015, 78) explains, there is a trend to design websites to be displayed in a different way on different devices, but not all “services residing on Internet servers are being configured to support user interactions across multiple devices.” For this research, we combine these various models to develop a framework for understanding the key factors that determine the nature of use of open practices that we use in this study.

The first step in our analysis is to create user profiles. These profiles are the most commonly occurring groups of characteristics in terms of each of the following: access location, access devices, and socioeconomic category.

The After Access Survey–2017 includes four possible places where people access the Internet: at home, a public space, a workplace, and a place of study (e.g., school, technical college, or university setting). We grouped those surveyed by the places of access. The first step toward creating the profiles is finding all the possible combinations of places from which people access the web, as shown in table 9.1. The most relevant combinations of the sample are in bold: all places mentioned (30 percent); all places except study (17 percent); home and public (14 percent); only home (13 percent); all places except workplace (10 percent), and home and work (7 percent). Each of the remaining combinations represents less than 5 percent of the sample.

The next step is distinguishing the relevant access combinations by sociodemo-graphic information, which includes gender,² position in the household (head, spouse, or child), age group, education level, main occupation, and socioeconomic level. Table 9.2 shows the percentages of access within a certain group (sociodemographic) for each proposed category (access combination). For example, the percentage shown for women indicates the extent of access from a particular combination of locations within the women’s sample (i.e., 32 percent of women access the Internet from all the places identified while 28 percent of men do so). This is telling us that women tend to specialize less than men do when choosing where to access the Internet. The categories that stand out from each group appear in bold in table 9.2, taking into account two criteria: the score for that category should be the highest in the category, at least by one percentage point, and should represent more than 10 percent of the whole sample that accesses the Internet. Sticking to the previous example, we highlighted the female gender because it is more than 1 percent higher than the male percentage and it represents more than 10 percent of the total sample. The same analysis is carried out for the rest of the categories. In the case that one group meets just one of the criteria, we also highlight a second group that meets both. That is why the under-age is highlighted in the age category as they compose less than 10 percent of the sample that uses the Internet.

Table 9.3 shows the main access location categories for various socioeconomic profiles.

In this second subsection, we apply the same procedure to identify combinations of devices: smartphone, tablet, and PC. Four relevant combinations were obtained, which in total represent more than 95 percent of the total sample, and these appear in bold in table 9.4. These combinations are access to all devices except tablet (48 percent), mobile phone only (34 percent), PC or laptop only (8 percent), and access to all devices (6 percent).

Within each access device combination, the percentages and the socioeconomic profiles are shown in tables 9.5 and 9.6, respectively. For the first relevant access combination, “All devices except tablet,” there are no significant differences between the female and male groups. Nevertheless, with regard to the rest of the categories, the following groups are highlighted: “Son/daughter,” “Young,” “Under-aged,” “Higher than secondary [education],” and “Student.” As expected, people under 25 tend to be mostly students. As millennials, this population has grown up with the Internet and is typically familiar with these access devices.

The increasing affordability of the smartphone is allowing more people than ever to access the Internet, to enter the digital economy, and to benefit from life-enhancing opportunities. In Latin America, there has been a tremendous increase in the consumption of mobile Internet data: it grew from 5 million gigabytes in 2010 to 956 million in 2017 (Ericsson 2017)—an exponential increase over the seven-year period. Perhaps unsurprisingly, then, individuals who responded “Mobile phone only” in the After Access Survey were users with the fewest years of Internet experience, these are further highlighted in our sociodemographic categories as “Female,” “Spouse,” “Elderly,” “High school dropout,” and “Nonactive.” From an inclusion perspective, it is important to understand what activities these users can realistically perform on a mobile phone alone.

Finally, the group that has access to all devices is characterized as being young or adult, and having education higher than the secondary level. Unsurprisingly, those with the most online skills and formal education are those who also have the greatest diversity of access.

The next section analyzes the effect of the type of access, both physical and digital, on the number of open activities carried out by an individual.

The second part of the analysis presented here focuses on the correlation between the profiles of access defined previously and the use of Internet, as defined by a set of activities related to education, work (job search or job-related activities), and engaging with government for services. Table 9.7 shows the types of activities, which we use as a proxy for open consumption practices.

In particular, we divide the practices into open and mostly open. In the case of open, we refer to tasks that do not imply any kind of cost (free). Mostly open activities are those that might have a minimal cost. For example, for education, the search and download of literature may involve a download cost, but the survey did not ask about these potential costs.

In total, we identified twenty-five tasks (outlined in table 9.7): nine in education, seven in terms of job searching or work, and nine for government services.

Figure 9.5 displays the number of activities that each individual carries out within each task group, related with the access location. The number of realized tasks by the agents is very low on average (in the sample, only three of the twenty-five tasks were used), but the standard deviation is relatively high (over 4). Moreover, the median value is zero, which means that more than 50 percent of Internet users in the sample did not engage in open activities. To facilitate the statistical treatment, the analysis is carried out in relative terms (i.e., the variable is normalized).³

For education-related activities, people who access from all the places mentioned or all places except work demonstrate a higher level of activity than the other groups. In contrast, people who access only from home or home and a public place tend to engage in the lowest amount of education-related activity.

Regarding job search or work activities, logically, people who work access from their workplace, and people who study access more from where they study. However, what is interesting here is that workers usually do some online education activities, while students do not tend to engage in work-or job search–related activities online. Moreover, on average, people engage in more educational than job search–related activities.

Open government activities are the least performed, with only two of nine activities reported on average (compared with two of seven for job search or work activities, and three of nine for education activities). Then, with regard to the differences in access places, the same pattern, as open job-search and work activities, remains. Additionally, the grouping for the total of tasks shows that those who access from all places do the most open activities online, whereas those who access only from home do the least.

Figure 9.6 shows the same analysis implemented in figure 9.5, but with differentiation by device.

There are observable differences in the number of tasks carried out by each group. For all cases, not surprisingly, the people who access from all the devices, as well as from mobile phone and PC, maintain a higher number of tasks. People who access the Internet only through a mobile phone perform the fewest tasks.

In the previous section, we performed a bivariate analysis of access location and type of device on the number of activities performed on the Internet, either relating to education, job search and work, or government. It is clear from this analysis that the access locations or devices used affect the types and number of open tasks in which users engage.

The results obtained in the previous section do not control for other relevant factors, such as general traits of the individual, her or his home, and other digital characteristics. This section, on the other hand, executes a multivariate analysis using probability and regression models, which allows us to observe the effect of the type of access (location and devices) while controlling for other personal characteristics. It is important to note that the regression includes country per-capita income, thereby controlling for the differences in wealth in each country.

Table 9.8 shows the first set of results: the impact of a series of variables on both the probability of using the Internet for the three tasks in question (columns 1 through 3), and the number of tasks realized (columns 4 through 6). Given the different nature of the dependent variables, we ran two regressions. In the first, we use a dichotomous variable that takes the value of 1 if any kind of open activity is performed, and 0 if not.⁴ The marginal effects results are shown in columns 1 through 3. For the second regression, the dependent variable is continuous and corresponds to a normalized tasks index.⁵ This set of results is shown in columns 4 through 6.

Next, we describe the results, starting with the effect of the general characteristics of the individual and her or his home: gender, age category, education level, type of occupation, socioeconomic level, rural area population, and local native language spoken.

With regard to gender, being female positively affects the probability of using the Internet to get an education, but negatively affects the number of and the probability of performing job searches and government-related tasks. With regard to the number of education-related tasks carried out, there is no significant effect, meaning that women do just as many educational activities as men.

Concerning age group, we include three categories in the regression analysis (young, adult, and elderly), and the coefficients and marginal effects must be interpreted as increase/decrease with respect to being under age (which is the omitted category). In almost all cases, except in educational activities, there is a significant advantage for all the age groups with respect to the omitted category in the regression. This makes sense because it is the adults who have more job search needs and interactions with government. Nevertheless, it is important to be cautious with this result because in general, the literature observes a negative impact of age on the acquisition of digital skills, with the elderly holding the most vulnerable position in that sense (Barrantes and Cozzubo 2015).

Education level shows a statistically significant effect on all the inspected variables, which brings out the importance of this variable in explaining digital skill appropriation and the realization of open activities. This is not a surprising result, as it extends the findings of other research that shows the importance of level of education for engaging online to include open activities.

When analyzing occupation, we include five categories (student, employed, employer, independent, and nonactive),⁶ and the coefficients and marginal effects must be interpreted as increasing/decreasing with respect to being unemployed (similar to the age group analysis). First, as expected, being a student has a positive impact on engaging in education-related tasks, both in the probability to engage in open education activities and in the number of tasks associated with education. Also, as expected, this same category has a negative impact on the realization of job search–related Internet activities, again both in probability and in number of tasks. This result confirms what our intuition suggests: students’ work-and job search–related needs are usually lower in relation to the rest of the sample; thereby, this group performs fewer job search–related activities on the Internet. With relation to government activities, students do not show a statistically significant difference from the omitted category (the unemployed).

On the other hand, the Employees, Independent, and Nonactive categories follow a similar pattern in terms of open job search–related activities. In particular, they show a statistically negative relationship with respect to the unemployed (the omitted category) in both the probability and the number of tasks. This is an expected result, however, because the unemployed have significantly greater needs to look for a job. Both employees and independent workers, by definition, have a current job, so they have less of a need to actively look for one.

When looking at educational and government-related activities, neither shows a significant level of correlation with any occupational category. This could be due to the general nature of the occupational categories. For example, the Employer category could be referring to jobs in very different economic sectors, thereby generating an average of zero.

To finish with the general characteristics of the individuals and household members, the socioeconomic level shows a positive, statistically significant impact in all cases, although it is of low magnitude. This finding is unsurprising, as it supports the literature that has reported an important effect that this variable has on the use of digital devices: usually, the poorest members of society are excluded from different Internet activities (Mendonça, Crespo, and Simões 2015; Galperin, Mariscal, and Barrantes 2014). Along the same line, a country’s gross domestic product (GDP) also plays a positive role in explaining the probability of engaging in open activities, although it is limited compared to the other explanatory variables considered.

Additionally, we included as independent variables rural location and local (Indigenous) languages spoken. For rural location, there is a significant negative impact in the case of job search–related activities for both the probability and number of tasks, and it is not significant in the rest of cases. For local (Indigenous) language speakers, there is no significant impact on any of the open activities shown in table 9.8. This could be due mainly to the heterogeneity of the surveyed population in the five countries analyzed. For example, in Paraguay, which contributes about 70 percent of the local language speakers in the sample, the Guarani language is not necessarily a reason for discrimination and exclusion, as it is in Peru, Guatemala, or Colombia.⁷

The rest of this chapter seeks to explain the effect of the ICT characteristics, locations, and types of access on the probability of engaging in different open activities. One finding is that the number of years of experience using the Internet is positively correlated with the probability and number of open tasks performed. However, this relationship changes with the age group. As Barrantes and Vargas (2017) found, Internet users at some point in time tend to converge with respect to the number of online engagement activities, but this convergence is conditional on their age group. In other words, no matter how much Internet experience mature adults have, they will always engage in fewer Internet activities on average than younger people. So, for certain populations, no matter how much Internet experience they have, they will not be able to increase their skills and use the web without some kind of intervention.

On the other hand, ICT social capital (e.g., the number of friends on Facebook or WhatsApp) is significantly and positively correlated to educational and job search–related open activities, in terms of both the probability to engage and the number of tasks performed. The previous relationship highlights the relevance of the informant’s social circle to engaging and appropriating online activities (Nam 2014; Norris 2003; Smoreda and Thomas 2001). However, in the case of open government activities, the previous relationship is not statistically significant. This could possibly be explained by the fact that is not enough to have friends who use a particular ICT device or a social media platform to engage in open government activities; supply being a key factor, or when government platforms are available, trust may be a key variable in using them. The survey did not allow us to capture the effect shown by Nam (2014), who finds that in engaging with government through open platforms, friends using this kind of service are key.

For the case of access devices and locations, the most important results are portrayed in table 9.8. For example, using all devices has the largest magnitude of significant impact in most cases. Regarding location, there is a similar trend—the informants who have access from all places tend to engage in more online activities than the rest, as shown in table 9.8 (recall that these users tend to be young and possess more than a secondary education). Additionally, some other regularities can be identified. Obviously, when study place is included within the set of access locations, a higher level of educational activities is observed, both in probability and in number of tasks, as was the case for the bivariate analysis. The opposite happens when we analyze Internet access from the workplace, where a positive effect of using the Internet for a job search is observed, both in probability and in number of tasks. Likewise, the locations associated with lower levels of both the probability of engaging in open activities and the number of activities are the home and public spaces. For a more detailed analysis, figures 9.7, 9.8, and 9.9 show a graphic analysis of the effects of the type of access on the probability of realizing different Internet activities.

In figures 9.7, 9.8, and 9.9, the average marginal effects of the type of access on the probability of carrying out an Internet activity are shown. It is necessary to note that the interpretation of these marginal effects is different from the interpretation of the discrete marginal effects, presented previously. The discrete effect shows how a dependent variable changes when a unit of the independent variable moves. On the other hand, the average marginal effect, presented in the figures, represents the probability that the dependent variable takes the value of 1, given a certain average level of the categorical variable. What the estimation shows, then, is the probability of realizing open activities on the Internet for different types of access. The calculation of these effects is done using the Probit model, presented in table 9.8.

Figure 9.7 shows how the various types of access affect the probability of using the Internet for educational purposes. The effects of using different devices are shown in the upper section, and the effects of locations are shown in the lower section. Regarding the different devices, all the relevant combinations that included PC/laptop showed higher probability to engage in open educational activities, except for “Only PC.” This could possibly highlight the mobile phone’s inability to adapt to the educational user’s needs. Educational platforms sometimes require advanced features like more advanced multimedia capabilities and more flexible software tools. While the mobile phone can be a more efficient and appropriate tool than a PC for some types of educational activities (Lee 2015), it is not yet the predominant mode for accessing educational opportunities online.

Concerning the location of access, the set of locations where a higher probability of carrying out educational activities is “All Places Combinations,” 0.5, followed by “Home, Study Place, and Public,” with 0.45. After that, the next combinations show similar probabilities.

Figure 9.8 shows the average marginal effects for the probability of carrying out open activities related to a job search. Similar to the education case, the combinations of devices that show the highest probability are “All Devices” and “Only PC,” highlighting again the necessity of a PC or laptop to engage in these kinds of activities. On the other hand, only mobile phone and other combinations (e.g., mobile phone and PC) present the lowest probability of carrying out open activities related to a job search among the device categories.

Considering different locations, similar to the education case, the “All Places” access alternative presents the highest probability to engage in job search online activities, while “Only Home” and “Home and Public” provide the lowest. Coupled with the results from education, this highlights the necessity of multiple location alternatives for users to engage in an open activity related to education or job search, which could also suggest that it is continuity in access that is important to users. It seems that the availability of accessing the Internet from multiple locations matters more for job search than other considerations.

Finally, figure 9.9 shows the results for open activities associated with government. Regarding devices, the person who uses “All Devices” has a slightly higher level of average probability of interacting with government than the other categories. The other levels are statistically the same. In the case of access locations, something similar happens because there is a high uncertainty in the effect of different kinds of access. In particular, the only significant effect is “All Places,” which could be explained by the reduced number of people who engage in open government activities (only around 12 percent of the sample).

In this chapter, we have tried to show how context, understood as the individual’s characteristics and her or his ways of accessing the Internet, affects the possibility of benefiting from the consumption of open activities, as proxied by use.

The analysis shows that both the location of the access (home, work, study place, or public space) and the access device (PC, mobile phone, laptop, and other) affect the probability and extent of engagement in various types of open activities. As one might expect, access from more locations or with more devices positively affects the number of open activities performed. This gives rise to two possible lines of explanation. The first relates to familiarity with the Internet, as proxied by diversification of devices and places of access, which matters more than device or place in explaining open consumption. The other is continuity of access, with people connecting wherever possible and preferring to change devices in order to remain connected, as mobile Internet access is not yet affordable in these Latin American countries (Agüero 2015; Viecens and Callorda 2016).

The econometric analysis confirmed this bivariate analysis, showing that the forms of access (from home or work, or using a PC or a smartphone) are correlated to the probability and extent of open activities. The data also showed that other characteristics, such as gender, age, education level, and socioeconomic status, significantly affect the number and type of open activities. For example, access from fewer devices, only from home, or characteristics associated with unemployment or nonactivity generate a lower number of open activities being realized. Furthermore, the locations and devices used can affect the number and type of online open tasks performed by the individual, as well as her or his capability to benefit from open activities. One interesting finding is that, while smartphones are touted as the solution to Internet access issues, a large part of the population does not take advantage of this access to engage in the open activities asked about in After Access 2017 survey.

The results emphasize the need for complementary policies addressed at including a larger portion of the population in the benefits that come from open activities. To exploit the benefits of open activities that Internet access can provide, affordable tariffs and devices are as important as the expansion of the infrastructure that is associated with high teledensity levels. Policies to improve home Internet access, which could begin with programs to encourage government employees to buy PCs or upgrade their mobile phones, could contribute to more people engaging in open consumption activities.

Likewise, the results suggest which locations (home, school, public spaces, or work-places) should be prioritized in ICT policies for each kind of digital activity. For example, if the government sets a goal to carry out a digital literacy policy for the use of open data, it would make more sense that the core of the policy focuses on workplaces rather than on homes, given the results of this study. Given the low levels of benefit obtained from people who access the Internet only from public places, another example of relevant policy could be to encourage a set of ICT activities related to job searching or even work, education, or government in public spaces. This way, the public spaces could be rediscovered and upgraded, and the appropriation of the ICT benefits of the most excluded groups could be improved.

Regardless of the ubiquity of Internet access, or teledensities well over 100 percent per country, data from these five Latin American countries show that the locations and forms of access, combined with people’s characteristics, matter. Reducing the consumption gaps of open activities is a pending task. By providing evidence on the factors associated with these gaps, this research hopes to have contributed to the design of sound policies to address them.

This chapter examines the relationship between open government data and social inclusion. Twenty-eight open data initiatives from the Global South are analyzed to find out how and in what contexts the publication of open government data tends to result in the inclusion of habitually marginalized communities in governance processes, such that they may lead better lives.

The relationship between open government data and social inclusion is examined by presenting an analysis of the outcomes of open data projects. This analysis is based on a constellation of factors that were identified as having a bearing on open data initiatives with respect to inclusion.

The findings indicate that open data can contribute to an increase in access and participation—both components of inclusion. In these cases, this particular finding indicates that a more open, participatory approach to governance practice is taking root. However, the findings also show that in the cases studied here, access and participation approaches to open government data have not successfully disrupted the concentration of power in political and other networks, and this has placed limits on open data’s contribution to a more inclusive society.

The chapter begins by presenting a theoretical framework for the analysis of the relationship between open data and inclusion. The framework sets out the complex relationship between social actors, information, and power in the network society. This is critical, we suggest, to developing a realistic analysis of the contexts in which open data activates their potential for transformation. The chapter then articulates the research question and presents the methodology used to operationalize those questions. The findings and discussion section that follows examines the factors affecting the relationship between open data and inclusion and how these factors are observed to play out across several open data initiatives in different contexts. The chapter ends with concluding remarks and an attempt to synthesize the insights that emerged in the preceding sections.

It is a truism to describe digital technology in general, and the Internet in particular, as having a pervasive and profound impact on the development of humankind. So much so that Castells and Himanen (2014) propose a new mode of development—informational development—that drives material wealth creation in the twenty-first century. While the creation of material wealth does not necessarily result in the betterment of the lives of all (Ravallion 2016), Castells and Himanen suggest that a synergistic relationship between informational and human development has the potential to underpin sustainable and egalitarian social progress. This comes with the recognition that development is not only about capital accumulation or social and economic progress, but about the widening of people’s choices (Haq 1998), the removal of unfreedoms (Sen 1999), as well as improvements in living conditions and the provision of basic needs (Nayyar 2003).

The relationship between informational and human development is dependent, however, on “social interests and power relationships” (Castells and Himanen 2014, 17). For example, if informational development drives material wealth creation, unequal access to information may create further imbalances and inequalities in people’s material conditions. As growth tends to favor higher-income groups to the detriment of those at the lower end of the income scale (Chen and Ravallion 2007; Ravallion 2016), the availability of information may empower the empowered without proper safeguards in place. Therefore, the role of government is critical, as it is government that determines the corrective and redistributive policies required to create the conditions for balanced, inclusive development (Castells and Himanen 2014; Kohli 2004; Chang 2003).

At the same time, however, governments are not immune to the politics of accumulation and distribution. Governments have used the mechanisms of the welfare state for political patronage and clientelism, often resulting in the double-edged sword of a disempowered citizenry on one side (through increased dependency on the state) and a government tightly coupled to the financial sector, to the exclusion of a significant portion of its population, on the other (Castells and Himanen 2014). This latter condition is made possible by the fact that knowledge-intensive, globally connected, and technologically advanced sectors of national economies operate in the ahistorical, placeless space of flows that characterize the new social order of the network society (Castells 2007, 2009). Castells and Himanen (2014, 14) comment, “This is, in fact, the structural basis for the growing inequality, polarization, and marginalization that characterize the situation in most countries in the world, despite high rates of growth in many areas of the planet.” Therefore, development strategies must contest and contend with the new structuration of society if it is to deliver equitable progress toward better lives for all.

Data are a valuable resource in informational development, prized and exploited by those small and highly advanced sectors of globally networked economies. Open government data [i.e., data free to be reused without any restriction (GovLab and Laurence 2014)] are also of value to these exclusive sectors, and bold claims have been made that open data are worth, in financial terms, $3 trillion per annum to the global economy (Manyika 2013). Yet it remains an open question who will benefit from this purported boost to the global economy.

Open government data, as a public rather than a private resource, are embedded in principles of universal access, participation, and transparency, and, according to advocates,¹ promise to restore trust between citizens and government by making the latter more accountable and inclusive (Horrigan and Rainie 2015; O’Hara 2012). Ideally, increasing transparency should empower citizens to participate in and hold public institutions to account for the distribution of benefits that support human development. Accountable governments, in turn, are more likely to put in place the kinds of policies required to restore a developmental balance in the form of a socially productive synergy between informational and human development.

Open data can play a role in restoring such a balance between informational and human development that proprietary data are precluded from playing. Common to all definitions of open data is the condition of universal access. Access is critical when there is an expectation for open data to lead to data use and impact of a more inclusive kind. If one dispenses with unfettered access to the data, one foregoes transparency and accountability (in governance); replicability, verifiability, and efficiency (in science); and, particularly in the case of open government data, equal opportunity and socioeconomic development. A core value held by many proponents of open data is that it is inherently inclusive as such, and (theoretically at least) it democratizes access to informational resources (Kitchin 2014).

But some have highlighted several issues about the theorized benefits of open data. As far back as 2010, Gurstein expressed concern about the false hope of open data automatically resulting in greater social inclusion. He took his cue from early evidence emerging from real-world outcomes: “By raising fundamental issues in understanding the societal aspects of e-governance, it highlights the need to replace politically neutered concepts like ‘transparency’, ‘efficiency’, ‘governance’, and ‘best practice’ with conceptually more rigorous terms that reflect the uneven terrain of power and control that governance embodies” (Gurstein 2010). Peixoto (2013) contends that open data as a form of transparency do not necessarily lead to accountability if political agency is absent. A survey of the contribution of open development by Bentley and Chib (2016) finds that the research literature does not support claims that open development projects nor open data initiatives typically result in the inclusion of poor and marginalized communities. One reason is that the benefits of openness are largely dependent on the capacities of users to exploit it as a resource for public benefit (Gurstein 2011; Davies and Bawa 2012; Canares and Shekhar 2016). How openness is structured and the contexts in which it is embedded are at least two further determinants of who benefits (Gurstein 2011; Johnson 2014). Research on open data, while cognizant of the lack of deep knowledge on the topic (Davies and Bawa 2012), has not explicitly been directed toward this fundamental concern raised from the outset.

One of the central challenges of open data, then, is to actualize the core value of inclusion. To do so requires an understanding of how and in what contexts open initiatives are inclusive, which open initiatives have been successful and which have been unsuccessful, who is included and who is excluded from the outcomes of these initiatives, and whether the inclusion made possible by open data has empowered those included to reprogram the network so that the benefits of development are more equitably distributed throughout society. To explore the relationship between open data and inclusion, this chapter deals with the following questions: Do open data contribute to inclusive development? More specifically, how, for whom, and in what contexts do open data contribute to inclusive development?

The literature on inclusion differs across disciplines. Inclusion, in economic development, occurs when growth is accompanied with falling inequality and decreasing levels of poverty (Piketty 2014; Ravallion 2016). In sociology, inclusion is an amalgamation of processes intersecting at both the individual and societal spheres that result either in the welcoming, acceptance, and provision for, or the ostracism, rejection, and nonacknowledgment of a person or a group of people (Allman 2013). In development studies, inclusion takes place when all people contribute to creating opportunities, participate in decision-making, and share in the social and economic benefits of development processes (Hickey, Sen, and Bukenya 2015). Castells, possibly attempting a supradisciplinary definition of inclusion, defines the concept as the systematic ability of individuals or groups to access the means for meaningful survival (Stalder 1998).

In this chapter, we anchor our view of inclusion in political participation—that is, whether individuals or groups of people enjoy equitable opportunities in shaping how they are governed and achieve and benefit from desired governance outcomes (Habermas 1996; Yuval-Davis 2011). Our approach to inclusion is political because it is inseparable from power, as highlighted by Gurstein (2010, 2011) and others [see, e.g., Arnstein’s citizen control (1969), Pretty’s self-mobilization (1995), and Hurlbert and Gupta’s adaptive governance (2015)]. Access and broad forms of participation are important stepping stones to inclusiveness but do not necessarily confer agency on or empower those being included; access and participation may simply result in a “voice without agency … [and] presence without politics” (Singh and Gurumurthy 2013, 186) for previously excluded communities.

Inclusion is meaningful if material wealth and nonmaterial benefits (such as dignity and health) accrue to those habitually excluded. For those benefits to accrue, though, access to information networks and participation in the decisions taken by powerful nodes² in those networks are necessary conditions, as networked information flows have become the primary setting for human agency (Castells 1996; Stalder 2005). As Castells (2017, 72) states,

There has been no economy and no society in the world in which wealth and power do not depend on information and knowledge. It has always been the absolutely critical matter for wealth generation and power generation. What has changed is … the information and communication technology revolution … with all its consequences: the ability to create organizational forms; the infrastructure and the rapidity of processing information, transforming it into knowledge; and using these transformations into knowledge to make actual changes in the production system.

Reilly and Smith (2013) argue that openness signifies more than access to technology; it entails changes to the patterns of development and the distribution of information, cultural production, and knowledge in the direction of what Walton et al. (2016, 4) refer to as “networked social morphologies.” According to the rules for such openness, traditional hierarchically controlled modes need to make way for spaces that support networked social arrangements. While open data promise universal access and democratized participation, any developmental strategy focused on “spaces for achieving openness” needs to be mindful of the space of information flows and the new networked social arrangements that this space introduces, including the unavoidable binary condition of inclusion in and exclusion from global networks of wealth and power (Castells 1996, 2009; Stalder 1998).

The social dynamics of networks draw attention to a critical missing element in the assumed revolutionary potential of open data. If access and participation are to have meaningful outcomes, the habitually and systematically excluded must be able to access and participate in networks with something of value to offer the powerful established nodes in networks where decision-making is concentrated and the program for the network is determined. Once included, the opportunity arises to attempt the reprogramming of the dominant logic of the network to redirect the distribution of benefits in a more equitable fashion (Castells 2009). This shaping is unavoidably a struggle for power (Gurumurthy and Singh 2016).

Castells (2007, 29; 2009, 42–47) identifies four types of power in the network society. Networking power is the power of social actors in global networks over those who are not included in these global networks. It operates along the binary conditions of exclusion and inclusion. Network power is the power exercised by those in networks (typically those at the source of the network’s formation) who prescribe the standards or protocols that determine the rules of acceptance. In this sense, those who execute network power coordinate social interaction in the networks according to rules for inclusion. Networked power is the relative power of social actors over other social actors in the same network; and the forms and processes of networked power are specific to each network. Network-making power is the power to program specific networks according to the interests and values of the programmers, and the ability to connect and ensure the cooperation of different networks, while defending the network from competition from other networks.

Programmers hold the power to program and reprogram networks; switchers hold the power to control the connecting points between networks. Crucially, within this formulation, people, collectives, and organizations are understood to be social actors who have the ability to disrupt the dominant power switches, and/or to push for change through what Castells calls “counterpower” (Castells 2007, 2009). Counterpower in the network society is exerted when social actors attempt to change the programs of specific networks or disrupt the switches that represent the dominant programmatic logic of the network.

In defining inclusion in this way, we acknowledge that we are setting the bar high by insisting that, in line with human potential paradigms, inclusiveness requires empowerment in the forms just described. Without inclusion culminating in those habitually excluded being in a position to negotiate different (and, hopefully, more equitable) decision-making outcomes by disrupting existing network power arrangements, we see little by way of meaningful returns from access and participation for those typically excluded from governance processes. At the same time, we acknowledge that a fully inclusive society is a utopian ideal. As Castells (1999) contends, exclusion is a structural condition of the network society and openness can be limited to certain constituencies. Nevertheless, open data make new power arrangements possible by disrupting previously closed information flows and by democratizing the monitoring of powerful actors to enforce the maintenance of a balance between informational and human development. This discussion will consider whether and how open data disrupt the existing distribution of power, and whether they activate different forms of counter-power in networks to yield a more inclusive, synergistic development.

The Open Data for Development (OD4D)³ initiative consists of a network of partners with the ambition of “empowering women and men around the world to achieve their own development goals and actively participate in decisions affecting their lives” (OD4D 2016, 44). OD4D has supported several initiatives to explore the connections between open data and development, with a particular focus on developing countries where the gaps between rich and poor, formal and informal economies, and skilled and unskilled are at their most acute.

A first round of analysis comprised an examination of OD4D projects that had formulated their own theory of change with regard to the inclusion of marginalized publics in decision-making as an outcome of the project.⁴ Projects were selected for analysis on the basis of the following criteria: (1) open data must have been published by the project or the project resulted in the use of existing open data; and (2) there were tangible, recorded outcomes that are attributable to the release or use of open government data. A total of fifteen projects were included in the analysis (see table 10.1).