Refugee crises, climate change, biodiversity losses, the threat of pandemics, significant gaps in sustainable infrastructure, and persistent fragility or armed conflict – the list of global challenges to social and economic development seems to get longer each day. Each successive challenge underscores the need for robust, flexible, and effective multilateral development banks and organisations (hereinafter referred to as multilaterals) to bring international partnership, financing, and expertise to bear on the world’s most layered and complex challenges.
Multilaterals have played an important part in human development for generations. The Bretton Woods institutions, created in 1944, sought to rebuild both Europe and the global economic order after the horrors of World War II. Over the following decades, the multilateral system developed a sharpened focus on poverty reduction and post-colonial reforms to drive social and economic improvements, including through the formation of regional development banks. During the 1950s, nascent recognition of the role of private enterprise as an enabler of development by multilaterals subsequently placed greater emphasis on investment in high-risk sectors and on countries to create jobs and raise living standards. The 1960s saw the advent of interest-free loans, advice, and grants intended to boost growth, reduce inequalities, and improve conditions in the world’s poorest countries. Following a stock market crash and global oil crisis in the early 1970s, multilaterals developed a broader recognition of their role in enabling better investment in the private sector across regions, and, by the 1980s, there was a focus on more and better insurance and guarantees to protect and spur foreign direct investments in developing countries. Toward the end of the last century, multilateral support for development began to broaden beyond work with just governments to direct work with, and support for, civil society, Indigenous communities, and non-governmental organisations. This work specifically looked to improve the quality and sustainability of social and economic development initiatives, to shine a light on corruption, and to give voice to demands for government transparency and accountability.1Throughout this period, the number of multilaterals has grown,2and over the last two decades, enabled by exponential changes in technology, they have evolved to include an increased focus on knowledge,34data, and, to some extent, open data.
There are a range of views concerning the comparative advantage of today’s multilaterals versus bilateral organisations (e.g. national aid agencies, foundations, and other donors). This chapter will not explore these perspectives in detail; however, for the purpose of this chapter, these comparative advantages include:
- A focus on strengthening social and/or economic development outcomes in middle-income countries (MICs), low-income countries (LICs), and/or fragile and conflict-affected states.
- The twin abilities to convene and work with a spectrum of stakeholders, including actors from across the public and private sectors at the national, regional, and global levels, and combinations thereof.
- In-depth technical knowledge across key sectors and in country contexts.
- The ability to mobilise and bring (or leverage) significant financing and technical support to bear.
With these criteria in mind, this chapter will briefly set out insights gleaned from interaction with several multilaterals that provide programmatic support for open data. This chapter is not meant to be exhaustive, but rather a snapshot of issues currently relevant for multilateral stakeholders in the open data landscape. The analysis that follows aims to discern relevant issues, reflecting on key lessons learned from effective and ineffective open data interventions, and to look forward to surface priorities and recommendations regarding the near-, medium-, and long-term opportunities and challenges for open data. While drawn from the experience of multilateral institutions, the hope is that this may be collectively useful for a wide range of stakeholders either currently engaged with open data or developing plans for the future.
The evolution of multilateral support for open data
Within the last decade, a number of multilaterals have become increasingly seized of the role and importance of open data for development. Several multilaterals have been in the vanguard of this great global disruption by opening their vast data holdings for free with the launch of their own open data initiatives. The World Bank was an early adopter in 2010,5and open data portals have since been launched by the African Development Bank (2013),6the Inter-American Development Bank (2015),7and the Asian Development Bank (2018),8among others. These open data platforms have provided access not just to edited summary statistics but to detailed datasets. In itself, this has marked a significant departure from the way multilaterals previously operated, and concepts such as “open by default” and “as open as possible” have since gathered much momentum throughout the international development community. This has resulted in a dramatic increase in users accessing and using data up and down the development value chain, including the increased use of data by staff within their own institutions.
Following their own experiments with open data, several multilaterals began to assist LICs and MICs with opening government data. The earliest case, World Bank support for Kenya’s open data initiative in 2010,910was swiftly followed by a range of work in countries across regions and at the subnational level, as well as the creation of the Open Data Readiness Assessment (ODRA) and its implementation in more than 45 countries as of 2017.11The Inter-American Development Bank, through the Open Government programme, has supported wide-reaching regional dialogue and knowledge sharing on open data, and has funded a number of thematic technical assistance projects, including work in 2015 on geospatial data12and open data for transparency applications.13The African Development Bank has supported countries in adopting open data practices by providing a shared data portal infrastructure and by bringing countries onto the “African Information Highway”, building on pre-existing statistical capacity development programmes.14
While not all multilateral development banks have embraced open data as a high-profile theme in their work, many have engaged as contributors to the International Aid Transparency Initiative (IATI), and a number have adopted specific thematic open data interventions, such as the European Bank for Reconstruction and Development’s support for Open Contracting (see Chapter 5: Development assistance and humanitarian action).15Other multilateral organisations have also been active in building networks around open data. For example, the Organisation for Economic Co-operation and Development (OECD) has provided backing for open data through regular studies, country reviews,16and expert meetings, and the United Nations Development Programme is a lead partner in the Open Data for Development Node in Europe and Central Asia (ODECA).17
Across these multilateral interventions, a range of approaches has been deployed, including technical assistance, capacity development, funding, the creation of knowledge products and tools, and the inauguration of international and global partnerships, communities of practice, and fora. Some of these interventions have helped to generate collaborations which have catalysed continuing initiatives, such as the International Open Data Conference (IODC). Others have been short-term interventions, such as online communities of expert consultants, set up to take advantage of the global convening power of multilaterals to rapidly identify potential partners for governments working on open data. Several examples of approaches are evident in the World Bank’s review of its own open data activities between 2012 and 2017.18
This forward momentum did not occur in a vacuum. Awareness of the role and value of open data for social and economic development blossomed relatively quickly through a range of important events and engagements around the start of the last decade. Through a number of reports, conferences, and meetings, multilaterals both were influenced by, and influenced, national, regional, and global open data activities, helping to develop a model for open data initiatives and establishing open data as a concept that more than just a few higher-income countries could explore.
It is also relevant to note the financial and practical support that multilaterals have given to the Open Government Partnership (OGP), which has become a key forum for furthering the open data agenda.19Governments, such as the United Kingdom (UK) and United States (US), for whom open data has been a domestic priority, have also used their positions as chairs of multilateral fora, such as the G8, to push forward open data activities, including the G8 Open Data Charter20and the launch of the Global Open Data for Agriculture and Nutrition (GODAN) initiative.21
As with other activities undertaken by multilaterals in the past, an initial focus on supporting governments in engaging with open data was soon followed by efforts to provide support to non-government actors from corporations and mobile network operators to civil society organisations, media organisations, and more. This support has focused both on the use of open data and on assisting organisations in opening their own data in support of social and economic development objectives.
Ultimately, in just a few short years, many thousands of datasets have been opened as free digital public goods available to all interested users. A multiplicity of open data-driven tools, services, publications, diagnostics, and analyses have been developed from this data, and open data policies and initiatives have been launched and have flowered in regions across the world. Taken together, this progress has undoubtedly enabled more equitable access to information and digital data by users around the globe and has supported more accountable and efficient public administrations, contributing to social development and economic growth.
Lessons learned: Supporting open data ecosystems
As described above, multilaterals have a unique perspective from which to reflect on lessons learned from the global open data landscape, in particular, how furthering the advance of open data can support social and economic development in the 21st century. This is particularly true with respect to a hard-won understanding of different operational successes and failures and how the fundamental building blocks of an effective open data ecosystem vary from one context to another. The experience of multilaterals has also resulted in the development of insights on how to support the broader development community to engage with open data through international cooperation, access to expertise and funding, and the provision of a range of public goods. These public goods, ranging from knowledge projects and data standards to software tools and shared algorithms, are resources that multilaterals are particularly well placed to provide. The following section provides an indicative synopsis of key learning points and opportunities for the future.
Open data maturity through alignment with other priority agendas
From the perspective of many multilaterals, open data as an issue has clearly reached policy maturity in part due to efforts by several multilaterals to integrate open data with other important agendas, such as the Sustainable Development Goals (SDGs) and national development plans. In addition to high-level discussions about the role and value of open data, multilaterals are likewise increasingly encouraging discussions about, and support for, open data at the local level and in sector-specific contexts, including work on standards and best practices in contracting,22statistics,23transport,2425urban resilience,26and a number of other areas. This “mainstreaming” of open data as a methodology in support of other agendas demonstrates an effort to move the agenda from dialogue to implementation. This is largely due to the recognition that while open data can be understood as a global public good, supporting the enabling environment for open data at different levels requires work focused on specific challenges, sources, and constituencies in the service of key social and/or economic development outcomes. Otherwise, securing engagement, and ultimately impact, is very difficult to achieve.
Multilaterals can provide the medium- and long-term support needed for social and economic development impacts from open data
While open data is showing signs of maturity in developed countries, success stories are still sparse in developing countries and the early sprouts of potential are very slow to grow. Advocates for, and practitioners of, open data initiatives in developing contexts need sustained financial, material, and human capital in order to bring together all the elements needed for open data impact.27Multilaterals are uniquely placed to provide such medium-to-long-term support by virtue of their clear mandates and deep engagement in developing contexts, including through their physical presence with on-the-ground experts who have established professional relationships among government and non-government actors. When used effectively, the funding modalities, presence, and physical infrastructure of multilaterals can enable authentically collaborative work, adding to momentum where it exists and ensuring context-sensitivity and tailored approaches that engender mutual trust to prioritise specific development outcomes. This does not necessarily include transplanting models that work in the developed world.
Sustaining political will requires a strategic approach
For long-term impact from open data, evangelism and short-term engagements are not enough. Institutionalising open data as a mechanism within larger sectoral initiatives requires ample political will up and down the value chain. Advocates and practitioners of open data need to discern what this means in practical terms and work to coordinate action. Multilaterals are often well-placed to support this, including leading or contributing to rigorous contextual analysis (drawing on in-country experts and practitioners) and mapping where key decision-makers and stakeholders fall on a spectrum of support/opposition. Building on this, multilaterals can help in identifying strategic entry points to achieve traction or can add to existing momentum over time. They are likewise well placed to deliver ODRAs and to further unpack the national and subnational enabling environment for open data.
Furthermore, it is clear that timing with respect to election cycles is important, as are connections between open data and related policy implementation and reform initiatives, such as Access to Information bills, management information system (MIS) implementations, procurement reforms, and OGP action plans. When multilaterals are able to draw on their cross-cutting work to ensure that the details and timing of open data interventions are strategically aligned, the foundations for impact can be strengthened. This requires the discipline to manage expectations and to balance a focus on relatively “low-hanging fruit”, and otherwise attractive or high-visibility approaches, with complementary medium-to-long-term approaches that might deliver more sustainable value.
Working with National Statistical Offices (NSOs) is key to improving data quality in developing countries
NSOs are an important stakeholder group, particularly for developing countries, where they are often the primary (if not the only) source of high-quality, official statistics, which could be published as open data under the right circumstances, especially if enabled by an authorising environment characterised by amenable leadership and receptivity to new ideas. NSOs are uniquely situated to encourage and support innovation when opportunities arise to reform or update statistical legislation. Several multilaterals have strong, long-standing partnerships with NSOs across developing and developed countries, as well as existing system-wide approaches to support the strengthening of NSO technical capacity and infrastructure through financing, grant programmes, and technical assistance. Multilaterals also play a key role in the development and dissemination of statistical methods and standards. These are all effective points from which to help encourage prioritisation and implementation of open data initiatives in developing countries and at the subnational level.
Multilaterals also need to recognise the importance of experimentation and adaptation to support NSOs to modernise in order to play their part in a fast-changing data ecosystem. At a minimum, multilaterals should encourage NSOs to recognise that they are joining a plurality of data producers. This involves collaborating with NSOs to explore new or additional roles and responsibilities, such as quality control and validation of open data, as well as becoming “infomediaries” themselves.28Ultimately, as NSOs assume a stronger coordination role across an expanding constellation of data producers, multilaterals can share relevant experiences on new collaborations and methods of partnership. This does not mean, however, that multilaterals have all the answers, nor should they impose top-down solutions. Adopting strategies where multilaterals fund successes without prescribing specific tasks is the way ahead and involves living with the risk that some investments may fail.
Increased focus is needed on data use
Multilaterals have not been particularly effective in promoting the sustained use of open data by governments and citizens. Open government data use by citizens in developing economies (for good governance or improved accountability) is still low. Recent analyses indicate that the majority of people around the world either will not or cannot use opened government data.2930For multilaterals to help realise the potential of open data for decision-making at various levels and across country contexts, the challenges of promoting data use need to be overcome. This necessitates a concerted effort among multilaterals and partners to make open data and civic technologies work for everyone (not just elites) through the creation of tools for a broader audience (including the most vulnerable groups) and the development of more meaningful and coordinated efforts to tackle the root causes of human, financial, and technical challenges to sustained open data use. It includes addressing ubiquitous capacity gaps in data literacy, inadequate salaries for data scientists and practitioners, insufficient resources to launch and maintain permanent open data platforms, constraints on opening data in local languages, difficulty maintaining standards and quality of data, and challenges in identifying actionable data to be opened. Other chapters in this volume explore these challenges in more depth. However, with their broad portfolios, multilaterals have the opportunity to address these issues systemically, from technical infrastructure and connectivity through to scaling up inclusive data literacy investments and working to make sure open data projects listen to the needs and priorities of non-elite, non-users, including the poor, marginalised, and chronically underserved. To date, no multilateral has set out a comprehensive vision for how work across all their programmes can contribute to the foundations for, culture of, and impacts from, open data. This should be a priority.
Careful engagement with private sector investment can bring value, but risks of proprietary data should be addressed
Several multilaterals have either increased or entirely focused their efforts on support for private sector markets to help address development challenges. The general rationale for doing so is to augment multilateral-led development by drawing in both private sector solutions and financing, which typically helps to complement scarce public financing, to help ensure developing countries can avoid high levels of debt and contingent liabilities. In the context of support for open data, multilaterals should focus investment on the broader open data ecosystem, which may necessarily implicate private sector solutions or resources for infrastructure. However, deliberation and attention must be paid to the inherent risks, particularly where data, systems, and civic applications may become proprietary, rather than open. For example, as explored in Chapter 12: Land ownership, it may be appealing to accept private financial support to improve land registers, but if this comes at the cost of proprietary ownership of the cadastre and the systems for maintaining it, the short-term gain may come with a long-term cost. We do not know, at present, how common this kind of private investment-for-data-ownership engagement is across developing countries or in multilateral portfolios. Likewise, the extent of public–private partnerships focused on data-sharing agreements among multilaterals is also unclear. However, it is not difficult to imagine that such engagements will occur.
In any case, multilaterals must be deliberate and vigilant. Each individual engagement necessitates careful analysis and understanding of any and all personal data privacy protection issues, thorough and thoughtful deliberation about any potential ethical issues arising from entering into data-sharing licensing arrangements with private sector actors, and the assurance that multilaterals carefully safeguard their integrity and role as honest brokers in all circumstances.
Looking ahead: The fourth industrial revolution
The world is at the threshold of what has been called the Fourth Industrial Revolution.3132Countries around the world are seeing exponential advances in artificial intelligence, disruptive technologies, and digital economies, which are occurring alongside mega-trends like climate change, globalisation, urbanisation, and the rise of social media. This combination of forces is rapidly transforming societies across the globe. In addition to the potential benefits of technological progress, there are also the spectres of structural job loss, income inequality, atomising of social consciousness, and loss of privacy – all factors with profound consequences for social and economic development across polities. Accompanying these trends is an increasing recognition of the need for new paradigmatic thinking about the relationships among governments, civil society, the private sector, and citizens, and the role of data, including open data, which underpins these relationships.
Moving forward, multilaterals could comprehensively integrate open data into what they do and how they operate in five key ways:
1. Leading by example. All multilaterals should make their own development and operational data “open by default”, subject to key legal and privacy considerations. This includes incorporating open data requirements into operational procurement processes, such that where products or services involving the creation, acquisition, management, or publication of data are funded by the multilateral, the default provision should be for any resulting data to be made available as open data and deposited in an open data catalogue or in-country equivalent. This will help to maximise the impact of open data, since multilaterals would be taking a consistent approach to applying open data principles to all relevant aspects of their institutional work. This would also vest multilaterals with even more institutional credibility to build, share, and support digital public goods, such as tools and infrastructure, that encourage the creation, management, and use of high-quality open data, international standards, and principles for open data.
As part of their leadership role, multilaterals should also endorse the International Open Data Charter and adopt positions consistent with it within other partnerships, activities, and business lines. Moreover, those multilaterals which are not yet signatories of IATI should sign up to ensure a comprehensive multilateral commitment to improving the quality, coverage, and timeliness of the data they publish in a consistent (IATI) format. They should likewise advocate for, participate in the development of, and adopt data standards which further the IATI vision for open data.
2. Mainstreaming open data across development support. Multilaterals should mainstream open data activities into their broader sectoral development operations. This would not only maximise the return on investments in data in general, it could enable more and better data discovery, access, and use, as well as provide a foundation for improvements across sectors, governance, and in citizen engagement. Building on a renewed recognition of the value of open source code and open knowledge, multilaterals should explore an “open by design” approach to programme development, considering how various aspects of openness can be built-in, rather than bolted-on to their interventions. This involves a step change from early modalities for open data support, where open data components were often included as a supplementary component of programmes that had already been largely designed, to an approach where open data principles are incorporated much earlier in the scoping of projects.
While adopting open data approaches should not necessarily be an immediate requirement for all work with multilaterals, and there will be cases where open approaches are not yet proven, multilaterals should always explore such approaches where appropriate and advise clients on their potential benefits. Key lessons should be parsed from across these efforts to support iterative learning and to scale successes.
3. Privacy. Conversations about open data led by multilaterals would benefit from a far stronger focus on privacy and data security issues. It has been pointed out that datasets with microdata across sectors (such as healthcare, education, transportation, criminal justice, property registration/housing, and voter registration, among other areas) may include individual records that may threaten individual privacy if released openly. Multilaterals should either facilitate or be part of the larger conversation to help governments find a public interest balance between privacy risks and maximising accessibility and use of open data.33
4. Fragile, conflict, and violence-afflicted states. The majority of the world’s poor live in fragile, conflict, and violence-affected states. Multilaterals have a particular role to play in these countries, where statistics and the generation of official data remain particularly neglected and under-invested areas. Due to a combination of factors, official statistics are often reported erratically, subject to quality issues, or not disseminated in a timely manner or a user-friendly format, limiting their role in decision-making. Truly open data in these contexts is rare, and efforts to provide systematic support for open data and to otherwise make that data which exists accessible for public analysis are often further complicated by the lack of political will and an enabling policy environment. Yet, accessible data may be key to coordinating action to address development challenges, and multilaterals must be careful to avoid contributing to a growing data-divide leaving these countries further disadvantaged. In particular, non-national holdings of government data (including those of multilaterals) may indeed be more robust than government holdings. When open, they can provide a useful stop-gap for analysis, capacity development, and development of knowledge products, although international provision of national data is not a sustainable long-term solution. As a result, multilaterals should explore how to shift resources to fragile, conflict, and violence-affected states, including resources for open data, comprehensive end-to-end data support, and data infrastructure which meets the particular needs of these states and, in particular, the poor and marginalised. This is especially critical in pursuit of global priorities, such as the SDGs and a range of poverty-alleviation priorities.
5. Investing in open approaches to monitoring, knowledge sharing, and research. Lastly, in order to help overcome the data use challenges outlined above, multilaterals should adopt a more coordinated, systematic approach to monitoring the use and impact of open data across their interventions, including the specific measurement of impact within sectors and for different types of open data. Effective impact measurement in the future cannot rely only on case studies and anecdotes. It will require well-resourced and rigorous research programmes that generate robust political economy analyses and data on returns on investment in open data interventions. Early open data advocacy made use of economic modelling to make the case for open approaches, and there is a need to bring back a stronger economics research focus on open data. Both the funding and convening power of multilaterals may be usefully deployed here.
If deployed effectively, multilateral contributions to the research base on what works in open data, and how open data works within specific sectors, have the potential to substantially advance the field over the next ten years, taking us far beyond the progress made already to date.
Every year multilaterals invest millions of dollars, considerable staff time, and political capital in supporting data-related projects, technical assistance, research, and knowledge sharing. It is only now, however, that they are starting to systematically track these data investments, laying the foundation for future research to understand how far they contribute to the open data landscape. While over the last decade, support for open data has become a small but established part of multilateral portfolios, the challenge for the decade ahead is to integrate open data approaches at the heart of wider data interventions to put “open by default” and “open by design” ideas into practice. This will require particular sensitivity to issues of privacy and security, as well as ongoing and in-depth research and learning to support constant improvement. However, as this chapter has explored, multilaterals bring a set of unique assets to this challenge and will undoubtedly continue to have a crucial role to play in taking open data into its next phase.