National Level Data Governance - A Review of Best Practices

This document presents how OECD countries are moving towards defining and implementing holistic public sector data governance practices at the national level. It discusses the main trends and challenges observed concerning data governance and proposes a public sector data governance framework drawing upon OECD best practices. The section then applies the model to briefly overview data governance practices across OECD member and partner countries.

1         Introduction

In the early 2000s, tech giants such as Facebook realised how digital platforms and the 24/7 connected citizen provided the ideal context to collect and re-use data for business purposes. The user data opened a window of opportunity to start selling data-driven products and services to any company and individual interested in designing ad hoc marketing and communication strategies – from businesses to politicians.

Data collected through multiple sources (from mobile phones to smart home devices) are now analysed to understand users better and target potential customers or service users. These insights may help organisations to drive citizens' choices, increase business revenues, influence public votes, or design and deliver better services. There is a plethora of technical solutions used for this purpose (e.g. artificial intelligence [AI], big data, customer relationship ma
nagement), which places the access to and sharing of data (EASD) as a precondition for data analysis techniques to help increase the value for companies and shareholders.

Since The Economist published the article, "The world's most valuable resource is no longer oil, but data" in 2017 (The Economist, 2017[1]), "data is the new oil" became the new buzzphrase and was sometimes abused and misunderstood by data enthusiasts. While this data-oil analogy aimed at increasing public awareness in response to raising data monopolies and controlled data flows, it also helped to stress how new technologies and data could help organisations make better decisions and increase business intelligence.

Still, while the discourse on "data as an asset" is well accepted nowadays, organisations, including the public sector, often fail to govern, manage and value data like the other relevant assets for their success. The neglect undermines the possibility of taking advantage of the opportunities brought by the "datisation of a huge amount of information that was previously intangible" (Chiesa, 2019[2]).

 Enabling the right cultural, policy, legal, regulatory, institutional, organisational, and technical environment is necessary to control, manage, share, protect and extract value from data. Yet, organisations from the public and private sectors often face legacy challenges inherited from analogue business models, ranging from outdated data infrastructures and data silos to skill gaps, regulatory barriers, the lack of leadership and accountability, and an organisational culture which is not prone to digital innovation and change.

New challenges have also arisen from citizens' data misuse and abuse, mainly by private sector organisations. The challenges are paired with the inability of governments to take proactive action, keep up with technological change, and understand the policy implications of data in terms of trust and fundamental rights.

Responding to these challenges requires a greater understanding of the structure and knowledge-sharing of how OECD countries address data governance in the public sector. The need for governance is well recognised by private sector actors but is only gaining traction in the government sphere.

This white paper presents a brief overview of how national governments across OECD member and partner countries increasingly address data governance or have worked on developing specific elements. The paper also presents a proposed public sector data governance model based on OECD good practices on data management and sharing within the public sector, open government data and digital government. While not exclusive, the elements presented in the data governance model can be beneficial and act as a guide.

2         The Case for Good Data Governance in the Public Sector

Good data governance can contribute to setting a shared vision, enhancing coherent implementation and coordination, and strengthening the institutional, regulatory, capacity and technical foundations to control better and manage the data value cycle, i.e. collect, generate, store, secure, process, share and re-use data, as means to enhance trust and deliver value.

Good data governance is imperative for governments that aim to become more data-driven as part of their digital strategy. It can help extract value from data assets, enabling greater data access, sharing, and integration at the organisational level and beyond and increasing overall efficiency and accountability. However, while the concept is not new, most OECD governments struggle to implement it.

The OECD has observed the following trends in the governance, management and sharing of public sector data:

a.      Data governance is increasingly relevant to data protection practices at the global scale, both exclusively and explicitly. Yet, a strong and unbalanced approach to data overprotection can reduce the value of data sharing, such as in the delivery of cross-border public services.

Recently, data misuse by private companies and increasing concerns from citizens about data management in the public sector have triggered government intervention to improve personal data protection (OECD, 2019[3]). As a result, the ethical and transparent use of data is now high on the political agenda.

Data flows have increased across organisations, sectors (e.g. business-to-government) and borders, adding another level of complexity to data governance in a globalised and interconnected world. Data governance is no longer limited to organisational boundaries but is a multinational concern resulting from cross-border data sharing.

In this context, international instruments such as the EU General Data Protection Regulation have sought to "give back to citizens the control over their data" (OECD, 2019[4]) and take cross-national action to prevent data misuse. The General Data Protection Regulation pushed the data protection agenda forward, thus underlying the need for common frameworks to ensure data protection across borders. Nevertheless, data overprotection can result from the misunderstanding of national and international regulations and drive change in policy approaches (e.g. from openness by default to "open if possible, protected if needed").

The global challenge at this stage is thus to ensure the right balance between free data flows and data protection, as stated by Japan's Prime Minister Abe during his keynote speech at the World Economic Forum in January 20192 (Japanese Government, 2019[5]).

b.      Data governance elements are often part of broader digital transformation policies. However, these components can be fragmented, thus reducing their whole-of-government value in terms of public sector integration and cohesion. A holistic data governance can help the government as a whole to join up.

While OECD countries have often defined elements relevant to public sector data governance in the context of digital government, open data, data management, and/or AI strategies and/or policies, these elements are often fragmented. In some scenarios, this disconnection is deeply rooted in the intricate governance arrangements supporting those policies (e.g., different public sector organisations leading these policies or lacking clarity in leadership and responsibilities), posing significant data integration and sharing barriers.

Holistic data governance can also help enable the Government as a Platform (one of the critical dimensions of a digital government). For instance, the development of standard but flexible data tools (e.g. data sharing platforms) provides solutions that can be re-used across the broad public sector. At a more technical level, fragmentation also results from legacy challenges regarding what the organisation generates. It controls the data and the impossibility of sharing and accessing those data in light of specific legal arrangements, leading to siloed policy and technical solutions that add to the impossibility of building an integrated and connected government. The lack of an overarching data governance model can lead to the proliferation or duplication of data standards and technical solutions for data sharing, thus hindering data interoperability across different organisations and sectors and affecting the possibility of integrating data, processes and organisations. It could also lead to multiple requests for citizens to provide the same personal data numerous times to the public sector unnecessarily.

A data governance framework must ensure the proper management of data through its entire life cycle (Ghavami, 2015[6]). For instance, in the past years, the open government data movement allowed for a more in-depth discussion of the need for strengthening data leadership and stewardship within the public sector. The governance discussion also opened a more technical discussion on improved data management practices, e.g. around the production, storing, processing and sharing towards higher data openness. Nevertheless, these elements were not understood as part of broader public sector data efforts connecting all stages of the data value cycle. Countries suddenly realised the value of cataloguing data for openness and discoverability purposes but have failed to acknowledge how these initiatives also had relevant policy benefits for productivity within the public sector.

On the other hand, in some OECD countries, a well-established culture of public sector efficiency led to the development of data registers to improve inter-institutional data sharing. Yet, this mind-set overshadowed the growing value of opening up government data and engaging and collaborating with external actors to solve policy challenges. As a result, those countries that once led the former e-government movement (with a strong focus on efficiency) lagged far behind those that doubled efforts to share and open up data to users to promote business and social innovation.

OECD countries such as Canada, Ireland, the Netherlands, the United Kingdom and the United States have moved towards the definition of overarching data strategies to build greater public sector cohesion and promote the integration of policies and tools.

These strategies comprise most, if not all, stages of the government data value cycle (from data production and its protection to data openness and re-use). Still, each stage requires specific arrangements, as they produce specific policy benefits (e.g., open data enables the use of data as a platform for greater user engagement and collaboration, and better data collection production practices can help reduce policy bias).

c.       Policy makers can misunderstand data governance as the exclusive responsibility of IT departments, but it also implies transformation and coherence of capacities, policies, regulatory frameworks, leadership and organisational culture. Therefore, more strategic public sector data governance approaches are needed.

The OECD has observed that a strong focus on technical issues as the primary outcome of data governance can misguide data-related policy decisions. For instance, by focusing primarily on the adoption of technological solutions such as application programming interfaces (APIs) and data standards (see the Overview of public sector data governance practices later in this chapter), rather than also enabling the adequate organisational, governance and cultural context to make those tools valuable to address policy challenges. All of these are critical elements of good data governance.

In some cases, OECD countries have invested resources to define strategic roles (e.g. data stewards, chief data officers) to support data governance by describing a more robust institutional fabric. Establishing these strategic roles can help scale and sustain policy implementation and build greater data maturity across the public sector (OECD, 2018[7]). The effort has occurred either in the context of data strategies or open data policies [e.g. Korea and the United States (see the Overview of public sector data governance practices later in this paper)]. However, in most countries, data leadership and/or stewardship are still misunderstood, thus confining data governance to the activities of the IT department and not as a factor that can help achieve policy goals through better data management and sharing practices.

d.      Public policies tend to overlook the benefits of data governance. There is a need for promoting data governance as a sublayer of policy arrangements. This can help extract value from data for a successful policy.

Good data governance supports public sector reform as a whole. In this light, its application is in line with core OECD principles and guidelines in areas such as regulatory policy (OECD, 2012[8]), digital government (OECD, 2014[9]), public procurement (OECD, 2015[10]), budgetary governance (OECD, 2015[11]), open government (OECD, 2017[12]), public sector integrity (OECD, 2017[13]), and public service leadership and capability (OECD, 2018[14]).

In best-case scenarios, most or some of the different elements of data governance (ranging from data strategies and institutional and regulatory frameworks to infrastructure and architecture) are nested within public sector digital transformation efforts and/or digital government policies. However, while policy and decision makers within line and coordination ministries (e.g. environment, transport, finance, public administration) increasingly recognise the relevance of "data as an asset" in their policy discourse, these policies often ignore the critical contribution of data governance to policy success. This context is not endemic to the public sector, for "today there is wide agreement that data is a critical asset [among businesses], but that doesn't always translate into taking the necessary actions to make that asset deliver real advantages" (Algmin and Zaino, 2018[15]).

The issue is particularly relevant in the context of cross-cutting public policies that require the sharing of, and access to, data from multiple public sector organisations for policy monitoring, compliance and evaluation purposes (e.g. public sector integrity, public budgeting, regulatory policy), or in the context of cross-sectoral data-sharing practices and governance arrangements (e.g. business-to-government data sharing) (see flexibility and scalability later in this section).

Public policies other than digital government can benefit from data governance as an underlying, yet mission-critical, element for policy success. This could be achieved when feasible by embedding different data governance elements in existing organisational and policy structures. By doing so, policy makers can enable the proper context and move from the overused discourse on data as an asset to the definition of an environment where data serves specific needs across the policy cycle.

e.      Good data governance does not happen in isolation. It benefits from adopting open, inclusive, iterative, collective and value-based approaches to its definition, implementation, evaluation and change.

Good data governance is not the responsibility of a small group of people. It should reflect the needs of a globalised, fast-paced, diverse, digitalised and inter-connected world. Public sectors must move away from closed and isolated ways of defining, implementing, monitoring and evaluating their data governance frameworks and tools.

Governments can benefit from adopting open, inclusive, iterative, collective and value-based data approaches when implementing their data governance initiatives. For instance, stakeholder engagement can help better identify data policy priorities and needs and assess the current context regarding data capability within the public sector. Iterative engagement can also help identify changing trends to take action and modify the rules and tools supporting data governance.

In addition, establishing partnerships with actors outside the public sector can help to:

·         Take advantage of private sector digital solutions to improve, streamline and modernise the public sector data infrastructure (e.g. cloud or Software-as-a-Service solutions)

·         Promote the publication of data produced by civil society organisations on government open data platforms or the publication of open government data on non-governmental data portals3

·         Support data sharing among multiple stakeholders from different sectors and increase data owners' control and decision power over the sharing and using their data to address common policy challenges [16](see Figure 1. Deploying data trusts as tools to pursue common value).

Good data governance also benefits from establishing a system of shared values and skills where all actors of the data ecosystem support and are responsible for policy success (e.g. data stewardship is shared among all relevant actors). At the same time, it implies defining and deploying a set of open and shared tools (e.g. open standards, APIs and algorithms) that can help promote integration within and outside the public sector.

2.1       Deploying data trusts as tools in the pursuit of common value

To accelerate the collection and sharing of data to harness artificial intelligence and other emerging technologies, governments and other organisations face the increasing need to explore and deploy tools for data management to protect data owners' rights while addressing common goals.

Data trusts add to data governance tools and build on long-standing legal trust frameworks applied to data management. They can also promote data sharing in areas where it is not currently happening. The Open Data Institute defines a data trust as a "legal structure that provides independent stewardship of data". Independent trustees are liable to make decisions about the data in the interests of the trust's beneficiaries, which may be other organisations, citizens, end consumers, or data users, by upholding laws and abiding by rules made when the data trust was set up.

In 2018, the United Kingdom launched its AI Sector Deal, a GBP 0.95 billion support package from the government and industry to keep the United Kingdom at the forefront of the artificial intelligence and data revolution. As part of the deal, the government committed to exploring data-sharing frameworks such as data trusts together with the artificial intelligence industry. The findings and recommendations of these pilots were published in April 2019 (Office for Artificial Intelligence, 2019[17]).

3         Developing a Common Framework for Public Sector Data Governance

While some countries have made advancements in clearly defining public sector data governance models, others have opted for a less strict approach where data governance is not explicitly acknowledged but implicitly takes place.

For instance, Luxembourg is working towards developing a data governance framework in the context of the recently adopted National Interoperability Framework. This work aims to take a more progressive approach that adopts the three core principles of digital-first, once-only, and transparency in the context of public sector data efforts. Luxembourg's National Interoperability Framework also sets objectives to promote open data, open standards and interoperability, machine-readable and linked data, APIs and open-source software in the public sector.

Yet, approaches to public sector data governance may vary in focus (e.g. a focus on technical governance aspects) or reach (e.g. specific data governance elements are available but dispersed).

For this reason, the OECD proposes a holistic model for data governance in the public sector to bring greater clarity and structure to the definition and implementation of the concept across countries. The model is based on the extensive OECD work on digital government and government data and additional research by the OECD Secretariat. Earlier versions of the model can be found in previous OECD digital government reviews, namely the OECD Digital Government Review of Norway (OECD, 2017[18]), the OECD Digital Government Review of Sweden (OECD, 2019[19]), the OECD Digital Government Review of Peru (OECD, 2019[20]) and the OECD Digital Government Review of Argentina (OECD, 2019[21]).

3.1       Data Governance Frameworks in the Public Sector - OECD Country Examples

3.1.1      New Zealand

The leading agency for government-held data in New Zealand (Stats NZ) developed a new and improved data governance framework for the New Zealand government. The framework is part of the agency's numerous efforts to promote better data management practices across the public sector and to leverage data as a strategic asset for decision-making. One of the central pillars of the framework is the adoption of a so-called "whole-of-data life cycle approach", meaning public bodies and employees are encouraged to think more strategically about the governance, management, quality and accountability of their data over the whole data life cycle (i.e. from the design and source of the data to its storing, publication and disposal).

Figure 1. New Zealand: Data governance framework[22]

3.1.2      Norway

As part of its work in developing Norway's national IT architecture, the Agency for Public Management and eGovernment created an information governance model that positioned public sector data management at the centre of the digital transformation of the Norwegian public sector. By placing data at the heart of the information governance model and by complementing it with strategic visions, policies, principles, standards and guidelines for better use of public sector data, public bodies in Norway have been given a rich set of tools to help leverage data as a strategic asset for decision making and re-use.

Figure 2. Norway: Information governance model[23]

3.1.3      Estonia

The data governance framework in Estonia is built on three core components – data source, handling and storage, and purpose – and stresses the importance of identifying and linking different data sources (e.g. private sector data, administrative data and census data) to varying types of data usages (e.g. policy analysis, research, operational), to ensure the proper handling and storage of data strategically.

Four main challenges (gathering, guarding, growing, and giving data) are crucial to creating a better data governance framework. These challenges cover a large section of the data value chain, from understanding data assets and establishing data governance principles to data processing, sharing and disseminating meta information.

3.2       Integration and Coherence

Good data governance promotes integration and systemic coherence and offers an everyday basis to use data to attain shared policy goals and foster trust. Ergo, the model intends to highlight the value of all organisational, policy and technical aspects for successful data governance. It identifies a range of (non-exclusive) data governance elements and tools and organises them into six groups. (see below and Figure 3. Data governance in the public sector)

These six groups are then arranged under three core layers of data governance (strategic, tactical and delivery) using the three traditional data governance categories as guidance (strategic, tactical and operational) as discussed and/or presented in Ghavami (2015[5]), DAMA Internal (2017[26]) and the BARC's 9-Feld-Matrix [see Grosser (2013[27]) and BARC (2019[28])]. The model is also based on additional research, including Ladley (2012[29]) and Sen (2019[30]):

3.2.1      Strategic Layer

See Figure 3 - (a) Leadership and vision

Some of the data governance elements in this layer include national data strategies and leadership roles. It is worth noting that the model considers data strategies as an element of good data governance. This argument rests on the fact that data strategies enable accountability and help define leadership, expectations, roles and goals. The strategic layer also highlights how the formulation of data policies and/or strategies can benefit from open and participatory processes, thus integrating actors' inputs from within and outside the public sector towards greater policy ownership.

3.2.2      Tactical Layer

See Figure 3 -  (b) Capacities for coherent implementation and (c) Legal and regulatory frameworks.

It enables the coherent implementation and steering of data-driven policies, strategies and/or initiatives. It draws upon the value of public sector skills and competences, job profiles, communication, coordination, and collaboration as instruments to improve the capacity of the public sector to extract value from data assets. It also highlights the value of formal and informal institutional networks and communities of practice as levers of public sector maturity and collective knowledge. This layer also comprises data-related legislation and regulations as instruments that help countries define, drive and ensure compliance with the rules and policies guiding data management, including data openness, protection and sharing.

3.2.3      Delivery Layer

See Figure 3 - (d) Integration of the data value cycle, (e) Data infrastructure and (f) Data architecture.

The delivery layer allows for the day-to-day implementation (or deployment) of organisational, sectoral, national or cross-border data strategies. It touches on different technical and policy aspects of the data value cycle across its stages (from data production and openness to re-use), the role and interaction of different actors in each stage (e.g., data providers), and the inter-connection of data flows across stages. Each stage is inter-connected but has specific policy implications with the expected outcomes. For instance, data-sharing initiatives (e.g. the production of good-quality, standardised and interoperable government data) can contribute to data re-use by external actors in later stages (e.g. as open government data). The adoption of technological solutions (e.g. cloud-based data-hosting services, APIs, data lakes) takes place in this layer, for it supports those policy goals defined in the strategic layer. It also relates, for instance, to the need for re-engineering legacy data management practices and processes or retrofitting and adapting legacy data infrastructures. Data interoperability and standardisation also take place at this level.

Figure 3. Data governance in the public sector[24]

The elements used to exemplify the plethora of policy instruments, arrangements, initiatives and/or tools that countries can use to deploy their data governance frameworks are not exhaustive. Thus, countries might adopt different data governance elements and tools that better fit their national context and public sector culture in line with the proposed three layers and the six underlying categories presented in the model.

For the analysis presented in this section, the data governance model explores practices at the national level (e.g. national data strategies, central data standards and national data-sharing platforms).

4       Flexibility and Scalability

The proliferation of data governance frameworks and tools in the public sector can hinder data integration and processes. Common policy goals (e.g. data protection) require coherent data governance frameworks, meaningful instruments (e.g. policies, regulations) and shared tools (e.g. data infrastructures, standards) that can help advance the cohesive deployment of data efforts in the public sector. Yet, the definition of a common data governance framework (from regulations and policy levers to standards and data federation tools and standards) should also allow for flexibility and scalability to avoid fragmentation, promote integration, and increase the adoption of good governance practices across organisations, levels of government, policy areas, sectors and borders.

This balance between adopting a structured approach and allowing for flexibility and scalability can help foster a common understanding, alignment and coherence of data efforts to support concerted actions, address shared policy challenges, and deliver joint policy results. Additionally, it can help to adjust the data governance model and its tools to specific contexts and respond to changing needs (e.g. anticipatory regulation) or ad hoc policy needs (e.g. different policy areas and stakeholders).

These arguments lay on the government as a platform dimension of digital governments. Thus, developing a coherent data governance framework enables public sector organisations to deploy and adopt standard data solutions and tools.

The different elements presented in the model and in this chapter address data governance from a national perspective (see the Overview of public sector data governance practices later in this chapter). However, the model is relevant in different contexts (inter-institutional, cross-border) where public sector data governance plays a crucial role in enabling data sharing and access.

The nature of the actors involved (the data ecosystem) can add to the complexity of the data governance environment as different actors have different needs and characteristics (e.g. sector, size) and differing digital and data maturities. However, the need for more excellent structure, flexibility, control, enforcement and compliance will also increase as the complexity of the data governance environment evolves, its purpose matures, and the needs of actors change, depending on whether it is implemented in a decentralised, federated or multinational context.

4.1       Organisational

At this level, data are shared across units, departments, and bodies within the same public sector organisation. Therefore, data governance can improve the management, sharing, and access to data within organisational boundaries. The need for a common data governance framework and shared data governance tools increases once actors external to the organisation join the data ecosystem.

4.2       Sector Specific or Policy Specific

Good data governance can also benefit a pool of public sector organisations that share common goals and mandates and produce and need to access share, or re-use standard datasets.

 

Earlier OECD efforts to promote good data governance in specific policy areas include the OECD Recommendation of the Council on Health Data Governance. It provides a set of principles to "encourage greater availability and processing of health data within countries and across borders for health-related public policy objectives while ensuring that risks to privacy and security are minimised and appropriately managed" (OECD, 2017[25]).

Examples of data governance initiatives in specific policy areas include the Geodata Strategy of the National Land Survey Authority in Sweden. The Geodata Strategy brought greater coherence and defined a set of common goals to foster the value of geodata for efficiency, innovation, competitiveness and the achievement of Agenda 2030 (Lantmäteriet, 2016[26]). The four pillars of the Swedish Geodata Strategy address different data governance elements, including interoperability, standardisation, openness and user engagement (OECD, 2019[27]).

The United Kingdom's Ordnance Survey provides another example of a maturing and more strategic sectoral data governance environment. In 2017, the Ordnance Survey (the UK national mapping authority) named its first chief data officer (Ordnance Survey, 2017[28]), and in 2019, it released its data strategy to continue delivering the benefits of sharing and opening accurate and quality mapping data for business impact (CIO UK, 2019[29]).

The Swedish and UK cases provide an organised and solid approach to opening up government data and highlight how the sharing of good-quality and trustworthy data requires taking action in the earlier stages of the data value cycle (e.g. data production) (see the Overview of public sector data governance practices later in this paper).

Another application case is the evidenced-based policy-making work carried out by the Japanese government. Japan has defined and implemented a solid, evidence-based and data-driven approach to improve the impact of policies and public services since 2017. This work draws upon data governance regulatory instruments published by the Japanese government, namely the Basic Act on the Advancement of Public and Private Sector Data Utilisation. For this purpose, the central government established a governance structure to ensure the coherent implementation of evidenced-based policy-making approaches across the broad public sector, including the establishment of a cross-ministerial council (which also benefits from the advice of external advisors) and the appointment of a director-general for evidenced-based policy making across all ministries at the central level. This case highlights the benefits of data governance and data for policy monitoring and adequate decision-making in the public sector (Fukaya, 2019[30]).

In Argentina, the Ministry of Justice developed a tool to improve the sharing of personal data in the context of judicial investigations using the central standard interoperability platform (INTEROPER.AR). The tool allows registered users (e.g. tribunals, prosecutors, courtrooms) to request data from and between those data registers connected to the interoperability platform (OECD, 2019[31]), speeding up data access and reducing the time to respond to citizens.

While in Argentina, there is a need to formalise data governance structures at the strategic layer, this case illustrates the potential scalability of the interoperability tool. For instance, its application can be expanded to other policy areas, including public sector integrity, as recommended in the OECD Digital Government Review of Argentina (OECD, 2019[32]) and the OECD Integrity Review of Argentina (OECD, 2019[33]). However, such an approach would require reinforcing the underlying data governance arrangements for public sector integrity while developing, implementing and/or adapting the specific rules and tools to respond to the ad hoc requirements of integrity policies.

This is particularly relevant as public sector integrity is a complex topic covering different areas with actors sharing and requesting standard data taxonomies for monitoring, reporting and/or auditing purposes (e.g. declarations of interest, gifts, open contracting data, beneficial ownership, budget data). Therefore, the importance of establishing a solid data architecture and infrastructure (technical layer) rests not on its benefits to inter-institutional data sharing but on the value of streamlined data-sharing practices to identify relationships between different stakeholders and reduce, monitor, control or address integrity risks.

4.3       Multilevel

Another level of complexity is added when data sharing occurs in a multilevel governance context. For instance, in federal models of government, the balance between central and local power impacts how the central government can access specific datasets owned and produced by local authorities.

In Mexico, a federal country, the central government developed the Open Mexico Network (Red Mexico Abierto, 2015-2017) to engage local governments in the central open data policy and facilitate the publication of open government data produced by local authorities on the central open data portal datos.gob.mx. For this purpose, the central government created a network of institutional contact points within public sector organisations at the state and municipality levels. This institutional fabric improved communication and coordination, but it also "ensured the efficient flow of tools and support provided by the federal government for the standardisation and publication of open government data" (OECD, 2018[34]).

Also, while central authorities can define overarching data quality standards, in practice, the responsibility for data quality falls on local governments, increasing the need to develop the proper controls to ensure that data are produced in line with central standards for policy monitoring purposes.

In Thailand, the former Ministry of Information Communication Technologies (now the Ministry of Digital Economy) designed a multilevel mechanism for reporting development data across all levels of government. While this initiative did not move forward, its architecture implied a complex data collection and sharing model, thus involving authorities at the local, provincial, departmental and ministerial levels under the leadership of the Office of the Prime Minister. This blend of actors, roles, and responsibilities requires strict data quality controls to ensure the data's quality, integrity, and trustworthiness across the entire data value cycle. Indeed, most of these authorities still face legacy challenges resulting from data fragmentation, duplicate standards, legal barriers and slow data-sharing processes, thus hampering the timely access to data for policy and decision-making (Wuttisorn, 2019[35]) and reinforcing the need for solid data governance.

4.4       Cross-sector

Common data governance frameworks contribute to effectively implementing cross-sector data collection, sharing and/or accessing initiatives. For instance, in regulatory compliance, business-to-government reporting practices can benefit from implementing common data governance structures and tools across all layers of the governance model.

In the Netherlands, the Standard Business Reporting (SBR)5 reduced the burden imposed on businesses in the provision of business information to local authorities and banks (SBR, 2019[40]). For this purpose, the SBR defined a shared public-private data governance framework, creating, among others:

·         A Steering Committee within the public sector defines the SBR's goals and programme of work, and a council decides the course of action, which benefits from insights from public and private sector actors. These elements reinforce the SBR's data governance strategic layer.

·         At the tactical layer, the SBR created a co-ordinator role to ensure coherent implementation of the programme. The SBR also developed a devoted platform where public and private sector actors can monitor and advise on implementing the programme.

·         At the delivery layer, the SBR standardised data definitions using a common data taxonomy defined by the Dutch government and streamlined and defined standard reporting processes. The digital government service (Logius)6 of the Dutch Ministry of the Interior and Kingdom Relations supports the technical aspects of the SBR.

Figure 4. Netherlands: Standard Business Reporting[36]

4.5       Cross-border

Increased data flows across borders demand more excellent government action to ensure data protection and ethical use, particularly citizens' data, when those are collected, processed and used by organisations from all sectors. The policy implications of cross-border data flows, both in terms of positive and negative benefits, are thus vast, and policy success requires the involvement of a plethora of actors at the global scale, from international organisations to businesses, data protection authorities and civil society organisations. OECD instruments such as the Guidelines on the Protection of Privacy and Transborder Flows of Personal Data (OECD, 2013[37]) have sought to bring greater coherence to cross-border data protection policies and initiatives across OECD member and partner countries.

 

Transborder data flows have specific implications for public governance and require more robust international data governance arrangements and coherent multinational action.

Reinforcing cross-border data governance can help monitor transnational infrastructure projects better and propel greater regional integration (for example, the Australia & New Zealand Infrastructure Pipeline, ANZIP). It also can support joint policy actions of governments to prevent, combat and tackle corruption at the regional level (e.g. by harmonising and enabling shared regulatory frameworks to allow data-driven evidence to be used for auditing purposes within and across governments, facilitating data access and sharing, etc.).

Shared data governance frameworks can also help to improve cross-border public service delivery. For instance, in 2013, Estonia and Finland agreed on a common agenda for developing a digital government to support the implementation of cross-border digital services in areas such as tax, health and education (OECD, 2015[38]). This enabled the deployment of Estonia's X-Road data-sharing platform8 (see the Overview of public sector data governance practices later in this chapter) in Finland. The interconnection of Estonia's and Finland's X-Road platforms in 2018 (VRK, 2018[39]) has also led to more significant, automated and secured cross-border data sharing, benefiting service users and supporting the future development of additional cross-border services in the region.

The success of the cross-border deployment of the X-Road between Estonia and Finland not only relies on technical issues; it also highlights the value of shared data governance policy structures at the strategic level. Drawing on the bilateral agreement signed in 2013, in 2017, Estonia and Finland agreed on the creation of the Nordic Institute for Interoperability and Solutions, which "ensure(s) the development and strategic management of the X-Road and other cross-border components for eGovernment infrastructure" (NIIS, 2019[40]).

Overview of public sector data governance practices at the national level across OECD member and partner countries

This section briefly overviews national practices across OECD member and partner countries. It presents evidence and data collected through different activities across the OECD under digital government when feasible. These include national peer reviews, cross-national reports, OECD surveys on digital government and open data, and work on the data-driven public sector.

4.6       Strategic layer

4.6.1      National data strategies

The importance of better managing, protecting and sharing data within the public sector is gaining traction across the OECD. In front-runner countries, this has led to the development of holistic national data strategies. These strategies are often nested within public sector digitalisation efforts. Notable examples include the United States' Federal Data Strategy, Canada's Data Strategy Roadmap for the Federal Public Service, the Government Data Agenda in the Netherlands and Ireland's Public Service Data Strategy.

For instance, the Dutch Government Data Agenda centres on the value of data as a tool to address policy and social challenges. The Dutch Ministry of the Interior and Kingdom Relations leads the implementation of the agenda, but both central and local governments are responsible for implementing it.

The agenda also "pays specific attention to the protection of public values and fundamental rights" (BZK, 2019[41]), thus including policy issues related to data ethics and algorithm transparency. The agenda integrates policy goals oriented to better data management in the public sector and the publication and re-use of open government data. The relevance of the public sector's organisational culture and knowledge-sharing for transformation change are also underlined, which is in line with the OECD approach to the digital transformation of the public sector [see, for instance, OECD (2019)[42]].

In Ireland, the central government recently launched the Public Service Data Strategy for 2019-2023.10 The Irish data strategy draws upon earlier data initiatives and policy instruments, including the National Data Infrastructure and the Open Data Strategy. The Irish data strategy clearly shows the need to bring a unified approach to public sector data initiatives and define shared principles, goals, and actions to support public sector cohesion (Office of the Government Chief Information Officer, 2019[43]).

4.6.1.1     Example -  The United States: Federal Data Strategy

In June 2019, the US government issued its Federal Data Strategy, which presents a ten-year vision to unlock the full potential of the country's federal data assets while safeguarding security, privacy and confidentiality. The data strategy focuses on three core principles (ethical governance, conscious design and a learning culture). It adds to several existing initiatives, policies, executive orders and laws that have helped make the United States a front-runner in terms of strategic management and re-use of government data over the past few decades.

To capture the linkage between user needs and appropriate management of data resources, the data strategy covers 40 practices that guide agencies throughout their adoption of the approach. To ensure coherent implementation of the strategy in its early phase, federal agencies must adhere to annual government action plans that include prioritised steps, time frames and responsible entities. A draft version of the 2019-2020 Federal Data Strategy Action Plan covers 16 steps seen as critical to launching the first phase of the data strategy vision, including the development of data ethics frameworks and data science training for federal employees.[44]  [45]  [46]

4.6.2      National Data Design Processes

The design process of national data strategies is also relevant. The OECD has observed, for instance, that late stakeholder engagement in the development of public sector digitalisation strategies can decrease policy awareness, clarity, accountability and ownership [see, for example, OECD (2019) [47]]. Early engagement can help identify policy challenges that would otherwise be ignored and bring relevant actors on board before implementing these strategies.

One relevant example in this respect is the open consultation process launched by the Department for Digital, Culture, Media and Sports in the United Kingdom to develop the UK National Data Strategy. In June 2019, the Department for Digital, Culture, Media and Sports conducted a public consultation to collect evidence and inform the development of its National Data Strategy. The data strategy development will be followed by a series of roundtables and testing exercises towards the publication of the final document in 2020 (DCMS, 2019[48]).

It is also important to mention that while countries are moving towards holistic policy approaches for public sector data practices, a vast group of OECD member and partner countries have had more focalised data policies for some time. Examples worth mentioning are the open data policies in countries like France, Korea, and Mexico (OECD, 2018) and well-grounded data register policies in Denmark, Italy, Norway, and Sweden.

The Danish Basic Data Registers programme, which dates back to 2013, has evolved from a strong focus on data-sharing practices within the public sector to a hybrid approach where core public sector data assets are shared for public access and re-use through a public data distributor. In addition, the programme emphasises integration, for it allows for public sector data access through web services and APIs (OECD, 2018[49]).

4.6.3      Leadership

The institutional governance model is also a core element of good data governance, providing clarity regarding leadership and accountability. However, it is essential to distinguish between political and administrative leadership roles. On the one hand, political leadership provides the high-level support needed to advance the policy agenda; however, political administration changes can lead to vacant positions, resulting in reduced political support for data policies. On the other hand, the leadership of top management positions helps to implement and steer policy design and implementation, thereby increasing the continuity and sustainability needed to deliver results across political terms.

That said, some countries have formalised leadership roles by attaching them to existent administrative structures. Relevant examples include the Government's Chief Data Steward in New Zealand, which the Chief Executive of Statistics New Zealand holds. The government's chief data steward is in charge of leading the country's data policy. New Zealand's case is also relevant in terms of policy accountability, as Stats NZ releases a quarterly dashboard "highlighting key deliverables for their data leadership role" under the Government's Chief Data Steward (Stats NZ, 2019[50]] ).

An earlier example is that of France's Administrateur Général des Données, created in 2014 (French Government, 2014[51]) and attached to the responsibilities of the head of the Etalab16 (the task force within the Office of the Prime Minister in charge of coordinating the open data and artificial intelligence policy in France). In Canada, the Data Strategy Roadmap for the Federal Public Service recommends the creation of a Government Chief Data Steward to "clarify roles and responsibilities around enterprise data leadership" (Government of Canada, 2018[52]).

Others, however, have followed different leadership models, which are less hierarchical and shared by various individuals and respond more to the culture within their public sector. This scenario is observed, for instance, in Nordic countries like Sweden, where the central government has opted for a more consensus-based leadership model in the form of a data taskforce composed of leading public sector agencies (OECD, 2019[53]).

In either scenario, transparent leadership is a precondition to help achieve policy goals (OECD, 2019[54]). It is also worth mentioning that in some cases, open data leadership positions might act as chief data officer (CDO) de facto, as in the case of Argentina (OECD, 2019[55]) and Mexico (OECD, 2016[56]).

4.7       Tactical Layer

Good data governance enables the coherent implementation of data policies. Yet, successful policy implementation relies on the intersection of different factors, ranging from establishing inter-institutional coordination bodies grounded in adequate institutional networks to capacity-building initiatives, collaboration and knowledge-sharing. Also, while complex, the availability of the appropriate regulatory frameworks (e.g. for data sharing, openness and protection) helps to create the right environment for policy instrumentation (e.g. by reducing burdens and barriers to data sharing) and in setting the rules for better-controlling data management practices in the public sector.

4.7.1      Steering and Policy Coordination Bodies

Examples of policy steering or coordination bodies include, for instance, Ireland's Data Governance Board, which was created to formalise a sustainable "governance structure for the Public Service, through which the development and implementation of data management standards, guidelines and activities can be overseen" (Office of the Government Chief Information Officer, 2019[57]).

In the United States, the draft action plan of the Federal Data Strategy foresees the creation of a Data Council within the White House Office of Management and Budget (OMB) by November 2019 (Federal Data Strategy Development Team, 2019[58]). While the OMB Data Council will help coordinate the Federal Data Strategy, it will also inform OMB's "budget priorities for data management and use" (idem). These bodies can also play an essential advisory role in ensuring that data strategies take a risk-management approach and anticipate and respond to policy challenges as they emerge. The Data Ethics Advisory Group in New Zealand provides an example.

 

4.7.2      Chief Data Officers, Institutional Networks and Data Stewardship

The need for more robust institutional networks and data stewardship in the public sector is also a growing priority for countries. This draws upon the urgency to enact a paradigm shift from a primarily technical perspective to one focused not only on compliance and control over data management and sharing practices but also on strategic goals and fostering a problem-solving approach centred on citizens.

As illustrated in previous OECD work on digital government and open data (OECD, 2016[59]; 2018[60]; 2019[61]), some countries have made a clear distinction between technical and strategic data roles in the context of open data policies as a means to emphasise that digital and data-driven transformation goes beyond mere technical aspects.

For instance, in Korea, the 2013 Act on the Promotion, Provision and Use of Public Data established the roles of "officers responsible for the provision of public data" and "data manager". Officers responsible for providing public data coordinate the central open data policy at the organisational level, translate its goals into clear actions, and liaise with other organisations for this purpose. Data managers are in charge of administrative and technical tasks, including compliance with data standards, data quality and data publication.

In the context of national data strategies, New Zealand's operational Data Governance Framework provides an exciting example where data stewardship is seen more as a skill to be built up among public officials rather than a formal role. This approach aims to embed "data accountability and best practice data management across all data-handling positions, to evolve beyond the need for traditional data governance roles (e.g. data custodians, data stewards)" (Sweeney, 2019[62]).

Figure 5. New Zealand: Data stewardship in the public sector (proposed model)[63]

In the United States, the 2018 Foundations for Evidence-Based Policy-making Act (signed into law on 14 January 2019) directs the head of each agency to "designate a non-political appointee employee in the agency as the chief data officer of the agency" (US Congress, 2019[64]). This is part of the provisions of the Open, Public, Electronic, and Necessary Government Data Act (OPEN Government Data Act), one of the Evidence-Based Policy-making Act (OECD, 2019[65]). These efforts contribute to building a more mature data governance ecosystem within the public sector, which can help to address potential sustainability risks across political administrations.

4.7.2.1     United States: Chief Data Officers

The provisions of the Open, Public, Electronic, and Necessary Government Data Act [66]describe the activities and role of institutional chief data officers as follows:

The chief data officer of an agency shall:

1.      1. Be responsible for life cycle data management

2.      coordinate with any official in the agency responsible for using, protecting, disseminating and generating data to ensure that the data needs of the agency are met

3.      manage data assets of the agency, including the standardisation of data format, sharing of data assets and publication of data assets per applicable law

4.      (…)

5.      (…)

6.      ensure that, to the extent practicable, agency data conform with data management best practices

7.      engage agency employees, the public and contractors in using public data assets and encourage collaborative approaches to improving data use

8.      support the performance improvement officer of the agency in identifying and using data to carry out the functions described in Section 1124(a)(2) of Title 31

9.      support the evaluation officer of the agency in obtaining data to carry out the functions described in Section 313(d) of Title 5

10.  review the impact of the infrastructure of the agency on data asset accessibility and coordinate with the chief information officer of the agency to improve such infrastructure to reduce barriers that inhibit data asset accessibility

11.  ensure that, to the extent practicable, the agency maximises the use of data in the agency, including for the production of evidence (as defined in Section 3561), cybersecurity and the improvement of agency operations

12.  identify points of contact for roles and responsibilities related to open data use and implementation (as required by the director)

13.  serve as the agency liaison to other agencies and the Office of Management and Budget on the best way to use existing agency data for statistical purposes (as defined in Section 3561)

14.  comply with any regulation and guidance issued under Subchapter III, including acquiring and maintaining any required certification and training.

4.7.3      Legal and Regulatory Frameworks

Regulation plays a vital role in data governance; thus, its implications are vast. Regulation helps define the rules to control the access to and sharing of data, promote openness, and ensure and enforce the protection of sensitive data. These instruments also help define and implement common data standards towards greater data interoperability and streamlined data-sharing practices. However, regulation can also be an obstacle to good data governance because the proliferation of fragmented instruments and uncoordinated efforts can hinder cross-institutional data integration and sharing. Taking an anticipatory approach can help identify risks and trends to implement the needed regulatory action to foster public sector readiness to change.

4.7.3.1     Anticipatory Innovation Governance

As digital transformation is speeding up and new and unforeseen risks emerge due to increased datafication, governments' ability to anticipate and act upon uncertain futures becomes increasingly essential. An important distinction between concepts has to be made:

 ·         Anticipation is creating knowledge – no matter how tentative or qualified – about the different possible futures. This may include, but is not limited to, developing not only scenarios of technological alternatives but also techno-moral (value-based) scenarios of the future (Nordmann, 2014[58]).

·         Anticipatory governance is acting on various inputs to manage emerging knowledge-based technologies and socio-economic developments while such management is still possible (Guston, 2014[59]). This may involve coordinated inputs from various governance functions (foresight, engagement, policy-making, funding, regulation, etc.).

·         Anticipatory regulation is a function of anticipatory governance that uses regulatory means to create space for various technology options to emerge for sandboxes, demonstrators, testbeds, etc. This requires iterative regulation and standards development around an emerging field (Armstrong and Rae, 2017[60]).

·         Anticipatory innovation governance is a broad-based capacity to actively explore options as part of broader anticipatory governance, with a particular aim of spurring on innovations (novel to the context, implemented and value-shifting products, services and processes) connected to uncertain futures in the hopes of shaping the former through the innovative practise (OECD Observatory of Public Sector Innovation (OPSI), 2019[61]).

Consequently, anticipation does not mean predicting the future; instead, it is about asking questions about plausible futures and then acting upon them by creating room for innovation (e.g., through regulation) or by creating mechanisms to explore different options in government itself. Most governments today do not have a system for anticipatory innovation governance (usually, mechanisms connected to the former are siloed under specific policy fields or functions, e.g. foresight). In the face of increased datafication, this is extremely important as choices made today regarding the ownership, interoperability, privacy and control of data will influence analytics and services built on the data that cannot be predicted or foreseen today. For the latter, different mechanisms to explore possible futures are needed. To this end, the Observatory of Public Sector Innovation has launched an Anticipatory Innovation Governance Project in which the OECD and leading countries will test different mechanisms for anticipation in practice.

4.7.3.2     Soft Law Instruments for Data Interoperability and Quality

Across OECD member and partner countries, examples of regulatory instruments related to data governance are vast. These instruments cover different policy issues, from sharing and interoperability to open government data.

In Brazil, the central government is advancing on the development of a new data-sharing decree which will help to improve clarity about the different levels of permitted access to government data [including full access, partial (restricted to only a few public sector organisations and bodies), protected data (the custodian defines data access rules)]. Data sharing is identified as one of the foundational principles of Brazil's Digital Governance Strategy towards more integrated public services, data openness and the creation of value for citizens (OECD, 2018[67]).

In the United Kingdom, the 2017 Digital Economy Act helped to bring further coherence and streamline data-sharing practices in the public sector with a resulting positive impact on citizens, including, for instance, by eliminating the vast range of previous legal gateways blocking data sharing among public sector organisations in the context of fuel poverty payment requests and payments (Roberts, 2019[68]).

Also, in 2017, Italy developed a set of technical regulations on the territorial data of public administrations in adherence with the EU INSPIRE Directive. Italy also created a national metadata catalogue to guarantee the discoverability and clarity of spatial data and related services. Italy has also implemented a more stringent regulatory framework to safeguard personal data and protect the public administration's data. These regulations, framed in the context of the Digital Administration Code and the Three Annual Plan for ICT in the Public Sector, define a set of security measures issued by the Agency for Digital Italy to evaluate and improve the digital security of the public sector.

Softer legal and regulatory instruments, such as codes of practice, recommendations or guidelines, often follow these instruments.

As described in the OECD Open Government Data Report (OECD, 2018[6]), countries have also made advancements in establishing suitable legal and regulatory environments for open government data. Recent examples include the 2016 Digital Republic Law (Loi pour une Republique Numerique) in France, the 2016 Basic Act on the Advancement of Public and Private Sector Data Utilisation in Japan and the 2017 Law for the Promotion of E-government in Germany (OECD, 2018[6]). Executive decrees on open government data are also available in Argentina, Brazil, Mexico and Peru.

4.7.3.2.1     Argentina: Guide for the Identification and Use of Interoperable (data) entities

As part of several efforts to bring order to data management and sharing practices within the Argentinian public sector, the National Direction of Public Data and Public Information published the Guide for the Identification and Use of Interoperable (data) entities[69]. The guide is an ongoing effort to ensure that public and private sector organisations can follow simple methods to generate, share and/or consume good-quality government data, therefore putting the data as a service vision in practice.

It guides how to produce simple identifiers for data that different public sector organisations produce but that at the same time are regularly shared among organisations (e.g. country > country_id). Consistent and increasing efforts have been underway since 2017 to ensure this core reference framework for government data is available through APIs.

4.7.3.2.2     France: The General Reference Framework for Interoperability

In France, the General Reference Framework for Interoperability [70]offers recommendations to promote interoperability across information systems within the public sector.

Following the rationale of the European Interoperability Framework, the French framework focuses on different levels of interoperability, setting standards for each level that are to be implemented by public sector organisations. Standards are therefore established for technical, semantic or syntactic interoperability to guarantee that public sector organisations, their dispositions and systems are as interoperable as possible:

·         Semantic interoperability refers to the meaning of different words, which often varies among public sector organisations. This interoperability aims to streamline the definition of words across public sector organisations to ensure agreement regarding the meaning of data exchanged and the context of the exchange.

·         Technical interoperability refers to data formats, exchange protocols, and the conditions and formats of storing these data. This interoperability ensures that data can be exchanged adequately among public sector organisations and in the correct format.

·         Syntactic interoperability is a subset of the technical interoperability as it focuses on the technical format data should have to be appropriately exchanged among public sector organisations.

4.7.3.2.3     Italy: White Paper on Artificial Intelligence

In March 2018, Italy published the White Paper on Artificial Intelligence[71]. The white paper recommends that all administrations ensure the quality and usability of the data they provide to ensure these data are used to test and refine artificial intelligence systems. Additional tools, modelled to fit the needs of the public administration concerning the use, interpretation and release of data, are available on the national data catalogue dati.gov.it and the National Guidelines for the Valorisation of Public Information Assets.

4.7.4      Skills: Capacity building, collaboration and knowledge-sharing

Public sector capacity, talent, and collective knowledge are core elements of good data governance in the public sector and broader public sector reforms, including digitalisation and innovation efforts. For this reason, OECD instruments such as the OECD Recommendations of the Council on Digital Government Strategies (OECD, 2014[72]) and on Public Service Leadership and Capability (OECD, 2018[73]), as well as the OECD Declaration on Public Sector Innovation, acknowledge their value as pillars of transformational and cultural change.

Building greater and systemic public sector capacity has different implications from a public sector data governance perspective, including:

·         Purpose (outcome): What is the purpose of (the policy issue)? Data governance must support the business strategy and achievement of the goals. This translates into the need for clarity regarding expected outcomes when implementing data governance initiatives. For instance, while closely related, a capacity-building programme specifically deployed to improve data sharing for public service delivery might differ from one that focuses on promoting ethics and values in the design of public sector algorithms.

·         Provider: Who provides support? In earlier stages of data-related initiatives, the support provided to public sector organisations will play a key role in increasing policy take up and awareness. In addition, this support can help build the right set of skills by providing training that will improve the capacity for implementation. For instance, in Mexico, the central government (2012-18) created the Open Data Squad as the government task force guiding public sector organisations in publishing open government data (OECD, 2018[74]).

·         Receiver: Who is the target of capacity-building activities? Good data governance in the public sector is translated into a different set of skills. It is needed for various groups of public officials, from political appointees or public managers to technicians. In Argentina, the Secretariat of Public Employment developed a series of skill development programmes that target different groups of public sector employees, for example, the Lideres en Acción programme for young officials and the Construyendo Nuestro Futuro programme for high-level public managers (OECD, 2019[75]). These initiatives complement those in place in the context of the activities of Argentinia's government innovation lab, LABgobar, which focuses on building more technical data skills.

·         Assessment: Which skills are needed to achieve the purpose? Better targeting capacity-building activities demands an evaluation of the current data capacity gaps. An example is the National Digital Skills survey conducted in New Zealand in 2017 to assess digital skills in the tech sector and across the government. The survey results informed the report Digital Skills for a Digital Nation and helped target capacity-building activities in the country (New Zealand Digital Skills Forum, 2018[76]).

·         Coherence: How can public sector organisations standardise the data skill needs? Using common job descriptions and frameworks improves coherence when attracting talent to the public sector and promotes inter-institutional mobility and career development. As referenced in earlier OECD work (OECD, 2019[77]), one of the most well-known frameworks for job descriptions in the digital and data domain is the United Kingdom's Digital, Data and Technology Profession Capability Framework.18

·         Mainstreaming: How to move from learning silos to collective knowledge? Digital and physical platforms and learning environments can help promote peer learning and knowledge sharing. They can also help identify, share and promote the mobility of existent talent within and across the public sector. Canada's cloud-based platform GCcollab is an example of a collaborative digital space that allows public servants, citizens, students and academics to exchange knowledge. The Canadian government has also created an agile model for public workforce mobilisation called Free Agents, which allows public servants to switch job positions across the government for short periods, depending on their skill set.

·         Openness and engagement: How can external talent and knowledge be leveraged? Good public sector data governance benefits from acknowledging that public sector organisations are not siloed entities in the data ecosystem. Open knowledge practices and partnerships with actors of the data ecosystem beyond the public sector, such as universities and entrepreneurs, can help build capacity within the public sector and attract talent when needed.

4.8       Delivery layer

The delivery layer integrates the set of processes, mechanisms and tools that allow for the operational implementation of data governance at a more granular level.

4.8.1      The data value cycle

The data value cycle is complex, for it is the crossroad of the most strategic and tactical aspects of the data governance (regulations, policies) with those that are more technical (e.g. the architecture and infrastructure supporting data management, sharing, access, control and re-use). For instance:

·         Different stages of the data value cycle call for various technical skills and roles (e.g. data custodians, data architects, data scientists). This draws on the different outputs resulting from data processing at each stage. Implementing training and capacity-building programmes at the tactical level supports the growing availability of these skills (see previous section).

·         Each stage of the data value cycle faces specific challenges that may require policy actions. For instance, bias can occur in the data collection stage, negatively affecting how policies are informed and the resulting interventions designed using those data as input. In the United Kingdom, the Department for Digital, Culture, Media and Sports has hosted events focusing on addressing the gender data gap (Roberts, 2019[63]), recognising that data on issues disproportionately affecting women are either never collected or of poor quality. To reduce gender bias in data collection, the UK government has developed a government portal devoted to gender data.21

·         The data value cycle is a continuum of inter-related, not siloed, stages, where different actors add value and contribute to data re-use. For instance, government initiatives focusing on producing good-quality data can contribute to greater interoperability, sharing and openness in later stages. In Argentina, the data-as-a-service approach aims at securing the production of good-quality and interoperable public sector data (OECD, 2019[3]). Using this tactic, the government facilitates publishing, sharing and re-using public sector data (including open data) by public entities and external consumers.

·         The data value cycle may reflect organisational processes resulting from legacy systems. Reassessing or re-engineering these processes is crucial to ensure that digitalisation and data-driven efforts contribute to transformation and avoid perpetuating inefficient processes in the digital world.

·         Data protection takes place (or should take place) across all different stages, with data custodians playing a pivotal role in ensuring the trustworthy and protected processing of the data. These officials should also manage risks of data corruption or data leaks (intentional or not) across the whole value chain, which can also have undesired effects on public trust.22

·         The data value creation process is not linear but cyclical (value cycle).23 The idea of a value cycle implies a shift in thinking from the value chain as a linear process to an iterative cycle that benefits from evolution and learning (Cordery, Woods and Collier, 2010[65]). When this rationale is applied to the data value chain, it reflects the whole policy-making process (from definition to implementation, evaluation and revision). It increases the impact of investments on sound data management practices, for data are continuously produced, analysed, shared, used and re-used to inform and evaluate policy.

4.8.2      National Data Infrastructures and Architectures

Some of the most technical aspects of data governance take place in the context of data infrastructure and architecture. These two elements can help advance data-sharing and management practices across institutions, sectors and borders and build the foundations for delivering public value (e.g. through better public service delivery).

 

Estonia's X-tee platform (X-Road until 2018) is one of the most well-known examples of a sound data-sharing infrastructure in the public sector. The development and deployment of the X-tee platform set the foundations for real-time data sharing between Estonian public sector organisations. Created in 2001, X-tee implies implementing a data federation model that helped build more effective, seamless and streamlined public services.

The value of the X-tee relies on its integrating role. Thus, it aims to provide a whole-of-government solution (government as a platform in practice) to enable the secure and authenticated sharing of data across previously siloed data sources. The use of the X-tee in Estonia is regulated by law, and public sector organisations willing to access or share data from or with other public sector organisations are obliged to use the X-tee tool. This helps avoid the proliferation of other data-sharing solutions in the public sector and promotes cohesion in Estonia. These efforts provided a cornerstone crucial to building a digital government, enabling integrated services and platforms within and outside the public sector, and increasing the benefits for citizens and businesses in the country. Also, the cross-border Estonian-Finnish X-Road platform model has been implemented in other countries, such as the Faroe Islands, Iceland, Japan and Kyrgyzstan (E-Estonia, 2018[78]).

Another example of the willingness of OECD countries to improve their national data infrastructure is the Data Federation Project in the United States.

4.8.2.1     United States: Data Federation Project

The US Data Federation Project aims to bring greater coherence to data federation practices in the US public sector to support policy decisions better, increase operational efficiencies, enable the diffusion of shared processes and infrastructures, foster an integrated government, and combat silos.

The proliferation of the different data federated models using various tools, processes, and infrastructure could, therefore, be prevented and gradually replaced with a single and scalable data federation model developed by the central government. This would follow a "government as a platform" approach. Thus, the overall goal is to build a shared tool for data federation that can be adopted across the public sector.

The project will draw on the collection of best practices regarding efforts to collect, combine and exchange data from disparate sources and across different public sector organisations and levels of government. In addition, it aims to establish data standards, offer guidelines, and deliver reusable tools such as automated aggregation to foster knowledge sharing across public sector organisations and effectively re-use government data from different sources.

4.8.2.2     Other Notable National Data Infrastructures and Architectures

Italy developed the National Digital Data Platform to improve its national data architecture and infrastructure. This platform offers big data solutions, including data lakes, to facilitate easy access to, sharing, and analysis of large volumes of raw and unstructured data from the public administration. It demonstrates an increasing understanding among governments of the need to design data infrastructures and architectures adapted to emerging new technologies, including artificial intelligence and machine learning. In the context of open data, the Italian data portal dati.gov.it also responds to the need for more robust collaborative data sharing within the public sector. It is based on the principle of a "federation of catalogues", which allows any public sector organisation to "feed" the data catalogue with periodic updates. The catalogue, therefore, also helps measure the outputs of the open data policy in terms of data availability.

Opportunities for greater openness and collaboration with external actors have emerged due to governments' demand for better and more efficient data-sharing infrastructures. For instance, in the United Kingdom, the Digital Marketplace project has brought external providers of digital solutions closer to the public sector by providing resources such as the G-cloud framework, which guides external suppliers of cloud-based services when delivering services to public bodies. Inspired by the UK model, Norway has launched a project to create a similar procurement platform for cloud-based services following its 2016 cloud-computing strategy.

Also, APIs are growing fast across OECD member and partner countries to integrate data, processes and organisations (including those outside the public sector) in real time. In Brazil, the central government's integration platform and API catalogue Conecta.gov29 allows public sector organisations to more efficiently and effectively share data between themselves, facilitating the implementation of the once-only principle (as defined by Brazilian law30 in 2017).

APIs are also being provided for public access in the context of open government data policies across different OECD countries, including Australia, Canada, Colombia, Denmark, France, Mexico, Portugal, Switzerland and the United Kingdom (OECD, 2018[79]).

As mentioned earlier in this chapter, the Nordic countries of Denmark, Norway and Sweden have all secured more robust policies for base data registers, enabling real-time sharing of public information within (and sometimes outside) the public sector. Realising the benefits of effective sharing of base registers, several other countries are starting to look at similar solutions. In Brazil, a new Data Sharing Decree [80] will include the creation of a citizen base register to improve the quality of citizen identification and biographical information and facilitate an end-to-end digital public service.

The need for greater data standardisation has also gained traction across OECD countries within the public sector and in cross-sectorial and international efforts to foster regulatory compliance, public sector accountability, integrity and citizen engagement. For instance:

·         As part of its quest to protect citizens' digital rights and personal information, the French National Commission on Information Technology and Civil Liberties (CNIL) created a standard on data protection governance[81], which comprises 25 technical requirements for private and public organisations managing personal data, to comply with the EU's General Data Protection Regulation. Singapore also provides technical guidelines for ethical data sharing between organisations, with its Trusted Data Sharing Framework [82] released in June 2019.

·         The XBRL [83] digital business reporting standard is an example of a data standard adopted by governments worldwide. It allows financial statements and reporting information to move rapidly, accurately and digitally between private and public sector organisations using a common reporting term language, simplifying regulatory compliance and business reporting. The XBRL standard is today used by governments in OECD countries such as Germany, Japan and the United States. The SBR project in the Netherlands (see flexibility and scalability earlier in this paper) is another excellent example of a country applying business reporting standards to cut red tape and improve regulatory compliance through digital solutions.

·         Partnerships such as the C5 (which groups Argentina, Colombia, France, Mexico, the United Kingdom and Ukraine) reflect cross-national efforts to spur the definition and implementation of coherent open contracting data practices. This includes adopting common international data standards such as the Open Contracting Data Standard[84], which offers a series of guidelines regarding the release of standardised, high-quality and reusable data and associated documents for each phase of a public contracting process. The recent partnership between the Open Contracting Partnership (leading the Open Contracting Data Standard) and the Infrastructure Transparency Initiative will help to pave further the way for the increased adoption of better data management and open data practices in the context of public infrastructure and enhance the quality of the Infrastructure Transparency Initiative's Infrastructure Data Standard.

____________________________

REFERENCES

[1] The Economist (2017), “The world’s most valuable resource is no longer oil, but data: Regulating the Internet giants”, The Economist, https://www.economist.com/leaders/2017/05/06/the-worlds-most-valuable-resource-is-no-longer-oil-but-data

[2] Chiesa, G. (2019), Technological Paradigms and Digital Eras: Data-driven Visions for Building Design, Springer International Publishing, http://dx.doi.org/10.1007/978-3-030-26199-3

[3] OECD (2019), Digital Government Review of Argentina: Accelerating the Digitalisation of the Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/354732cc-en

[4] Ibid

[5] Japanese Government (2019), Toward a New Era of “Hope-Driven Economy“: The Prime Minister’s Keynote Speech at the World Economic Forum Annual Meeting, Prime Minister of Japan and His Cabinet, Tokyo, https://japan.kantei.go.jp/98_abe/statement/201901/_00003.html

[6] Ghavami, P. (2015), Big Data Governance: Modern Data Management Principles for Hadoop, NoSQL & Big Data Analytics, CreateSpace Independent Publishing Platform

[7] OECD (2018), Open Government Data Report: Enhancing Policy Maturity for Sustainable Impact, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264305847-en

[8] OECD (2012), Recommendation of the Council on Regulatory Policy and Governance, OECD Publishing, Paris, https://doi.org/10.1787/9789264209022-en

[9] OECD (2014), Recommendation of the Council on Digital Government Strategies, OECD, Paris, http://www.oecd.org/gov/digital-government/Recommendation-digital-government-strategies.pdf

[10] OECD (2015), Recommendation of the Council on Public Procurement, OECD, Paris, https://www.oecd.org/gov/public-procurement/OECD-Recommendation-on-Public-Procurement.pdf

[11] OECD (2015), Recommendation of the Council on Budgetary Governance, OECD, Paris, https://www.oecd.org/gov/budgeting/Recommendation-of-the-Council-on-Budgetary-Governance.pdf

[12] OECD (2017), Recommendation of the Council on Open Government, OECD, Paris, http://acts.oecd.orgRECOMMENDATIONPUBLICGOVERNANCE

[13] OECD (2017), OECD Recommendation of the Council on Public Integrity, OECD, Paris, http://www.oecd.org/gov/ethics/recommendation-public-integrity

[14] OECD (2018), Recommendation of the Council on Public Service Leadership and Capability, OECD, Paris, http://www.oecd.org/gov/pem/recommendation-on-public-service-leadership-and-capability.htm

[15] Algmin, A. and J. Zaino (2018), Trends in Data Governance and Data Stewardship: A 2018 DATAVERSITY Report, DATAVERSITY Education, LLC, http://content.dataversity.net/rs/656-WMW-918/images/Trends%20in%20Data%20Governance%20and%20Stewardship_FinalRP-Graphs.pdf

[16] See the work on data collaboratives led by Govlab in the United States: https://datacollaboratives.org

[17] For more information see: https://www.nesta.org.uk/report/renewing-regulation-anticipatory-regulation-in-an-age-of-disruption

[18] OECD (2017), Digital Government Review of Norway: Boosting the Digital Transformation of the Public Sector, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264279742-en

[19] OECD (2019), Digital Government Review of Sweden: Towards a Data-driven Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/24131962

[20] OECD (2019), Digital Government in Peru: Working Closely with Citizens, OECD Publishing, Paris, https://doi.org/10.1787/0c1eb85b-en

[21] OECD (2019), Digital Government Review of Argentina: Accelerating the Digitalisation of the Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/354732cc-en

[22] Sweeney, K. (2019), “An operational data governance framework for New Zealand government”, https://statsnz.contentdm.oclc.org/digital/collection/p20045coll1/id/2657

[23] OECD (2017), Digital Government Review of Norway: Boosting the Digital Transformation of the Public Sector, https://doi.org/10.1787/9789264279742-en

[24] OECD (2019), Digital Government Review of Argentina: Accelerating the Digitalisation of the Public Sector, https://doi.org/10.1787/354732cc-en

[25] OECD (2017), Recommendation of the Council on Health Data Governance, OECD, Paris, https://www.oecd.org/els/health-systems/health-data-governance.htm

[26] Lantmäteriet (2016), The Swedish National Geodata Strategy 2016-2020: Well Developed Collaboration for Open and Usable Geodata Via Services, Lantmäteriet, https://www.geodata.se/globalassets/dokumentarkiv/styrning-och-uppfoljning/geodatastrategin/national_geodata_strategy_2016-2020.pdf

[27] OECD (2019), Digital Government Review of Sweden: Towards a Data-driven Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/24131962

[28] Ordnance Survey (2017), Ordnance Survey appoints new chief data officer, 28 June, https://www.ordnancesurvey.co.uk/about/news/2017/carolinebellamy_chief_data_officer.html

[29] CIO UK (2019), Ordnance Survey Chief Data Officer Caroline Bellamy reveals data strategy, 20 February, 2019, https://www.cio.co.uk/cio-interviews/ordnance-survey-chief-data-officer-caroline-bellamy-explains-strategy-3692557/

[30] Fukaya, T. (2019), “Is evidence contributing to public accountability? Evidence from Japan”, presentation at the OECD Expert Meeting on Standards of Evidence, Ministry of Internal Affairs and Communications, Japan

[31] OECD (2019), Digital Government Review of Argentina: Accelerating the Digitalisation of the Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/354732cc-en

[32] Ibid

[33] OECD (2019), OECD Integrity Review of Argentina: Achieving Systemic and Sustained Change, OECD Publishing, Paris, https://doi.org/10.1787/22190414

[34] OECD (2018), Open Government Data in Mexico: The Way Forward, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264297944-en

[35] Wuttisorn, P. (2019), Open and Connected Governance in Thailand, Office of the National Digital Economy and Society Commission

[36] Groenveld, B. (2019), Standard Business Reporting (SBR), Dutch Ministry of the Interior and Kingdom Relations

[37] OECD (2013), Guidelines on the Protection of Privacy and Transborder Flows of Personal Data, OECD, Paris, https://www.oecd.org/internet/ieconomy/privacy-guidelines.htm

[38] OECD (2015), OECD Public Governance Reviews: Estonia and Finland: Fostering Strategic Capacity across Governments and Digital Services across Borders, OECD Public Governance Reviews, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264229334-en

[39] VRK (2018), Finland’s and Estonia’s Data Exchange Layers Connected to One Another on 7 February: The Rapid Exchange of Information Between the Countries Is Now Possible, Population Register Centre, Helsinki, https://vrk.fi/en/article/-/asset_publisher/suomen-ja-viron-palveluvaylat-liitetty-yhteen-7-2-tietojen-nopea-ja-luotettava-vaihto-maiden-valilla-nyt-mahdollista

[40] NIIS (2019), Nordic Institute for Interoperability Solutions: History of the Institute, Nordic Institute for Interoperability Solutions, https://www.niis.org/history

[41] BZK (2019), Data Agenda Government (Data Agenda Overheid), Ministry of the Interior and Kingdom Relations, https://www.nldigitalgovernment.nl/wp-content/uploads/sites/11/2019/04/data-agenda-government.pdf

[42] OECD (2019), Digital Government Review of Sweden: Towards a Data-driven Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/24131962

[43] Office of the Government Chief Information Officer (2019), Public Service Data Strategy 2019-2023, Government of Ireland, https://www.osi.ie/wp-content/uploads/2018/12/Public-Service-Data-Strategy-2019-2023.pdf

[44] Executive Office of the President (2019[48], Federal Data Strategy: A Framework for Consistency, https://www.whitehouse.gov/wp-content/uploads/2019/06/M-19-18.pdf

[45] Federal Data Strategy Development Team (2019), 2019-2020 Draft Federal Data Strategy Action Plan, US Government, Washington, DC, https://strategy.data.gov/action-plan

[46] 019-2020 Draft Federal Data Strategy Action Plan, https://strategy.data.gov/action-plan

[47] OECD (2019), Digital Government Review of Sweden: Towards a Data-driven Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/24131962

[48] DCMS (2019), National Data Strategy: Guidance, Department for Digital, Culture, Media and Sports, London, https://www.gov.uk/guidance/national-data-strategy

[49] OECD (2018), Open Government Data Survey 3.0, OECD, Paris

[50] Stats NZ (2019), Data Leadership Quarterly Dashboard, New Zealand Government, https://www.data.govt.nz/about/government-chief-data-steward-gcds/data-dashboard

[51] French Government (2014), Décret n° 2014-1050 du 16 septembre 2014 instituant un administrateur général des données, French Government, Paris, https://www.legifrance.gouv.fr/affichTexte.do;jsessionid=?cidTexte=JORFTEXT000029463482&dateTexte=&oldAction=dernierJO&categorieLien=id

[52] Government of Canada (2018), Report to the Clerk of the Privy Council: Data Strategy Roadmap for the Federal Public Service, Goverment of Canada, Ottawa, https://www.canada.ca/content/dam/pco-bcp/documents/clk/Data_Strategy_Roadmap_ENG.pdf

[53] OECD (2019), Digital Government Review of Sweden: Towards a Data-driven Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/24131962

[54] OECD (2019), Digital Government Review of Argentina: Accelerating the Digitalisation of the Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/354732cc-en

[55] Ibid

[56] OECD (2016), Open Government Data Review of Mexico: Data Reuse for Public Sector Impact and Innovation, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264259270-en

[57] Office of the Government Chief Information Officer (2019), Public Service Data Strategy 2019-2023, Government of Ireland, https://www.osi.ie/wp-content/uploads/2018/12/Public-Service-Data-Strategy-2019-2023.pdf

[58] Federal Data Strategy Development Team (2019), 2019-2020 Draft Federal Data Strategy Action Plan, US Government, Washington, DC, https://strategy.data.gov/action-plan

[59] OECD (2016), Open government data review of Mexico : data reuse for public sector impact and innovation., OECD, https://doi.org/10.1787/24131962

[60] OECD (2018), Open Government Data Report: Enhancing Policy Maturity for Sustainable Impact, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264305847-en

[61] OECD (2019), Digital Government Review of Argentina: Accelerating the Digitalisation of the Public Sector, OECD Publishing, https://doi.org/10.1787/24131962

[62] Sweeney, K. (2019), “An operational data governance framework for New Zealand government”, Stats NZ, Wellington, https://statsnz.contentdm.oclc.org/digital/collection/p20045coll1/id/2657

[63] Sweeney, K. (2019), “An operational data governance framework for New Zealand government”, https://statsnz.contentdm.oclc.org/digital/collection/p20045coll1/id/2657

[64] US Congress (2019), H.R.4174: Foundations for Evidence-Based Policymaking Act of 2018, US Congress, Washington, DC, https://www.congress.gov/bill/115th-congress/house-bill/4174/text

[65] OECD (2019), Digital Government Review of Argentina: Accelerating the Digitalisation of the Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/354732cc-en

[66] US Congress s (2019), H.R.4174: Foundations for Evidence-Based Policymaking Act of 2018, https://www.congress.gov/bill/115th-congress/house-bill/4174/text

[67] OECD (2018), Digital Government Review of Brazil: Towards the Digital Transformation of the Public Sector, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264307636-en

[68] Roberts, S. (2019), Data in UK Government, UK Department for Digital, Culture, Media and Sport, London

[69] Digital Government Review of Argentina: Accelerating the Digitalisation of the Public Sector, https://doi.org/10.1787/354732cc-en

[70] Direction Interministérielle du Numérique et du Système d’Information et de Communication de l’État (2015), Référentiel Général d’Interopérabilité: Standardiser, s’aligner et se focaliser pour échanger efficacement, http://references.modernisation.gouv.fr/sites/default/files/Referentiel_General_Interoperabilite_ V2.pdf

[71] Italy: AGID (2018), White Paper on Artificial Intelligence at the service of citizens, Available at: https://ia.italia.it/assets/whitepaper.pdf

[72] OECD (2014), Recommendation of the Council on Digital Government Strategies, OECD, Paris, http://www.oecd.org/gov/digital-government/Recommendation-digital-government-strategies.pdf

[73] OECD (2018), Recommendation of the Council on Public Service Leadership and Capability, OECD, Paris, http://www.oecd.org/gov/pem/recommendation-on-public-service-leadership-and-capability.htm

[74] OECD (2018), Open Government Data Report: Enhancing Policy Maturity for Sustainable Impact, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264305847-en

[75] OECD (2019), Digital Government Review of Argentina: Accelerating the Digitalisation of the Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/354732cc-en

[76] New Zealand Digital Skills Forum (2018), Digital Skills for a Digital Nation: An Analysis of the Digital Skills Landscape of New Zealand, New Zealand Digital Skills Forum, https://digitalskillsforum.files.wordpress.com/2018/01/digital-skills-for-a-digital-nation-online.pdf

[77] OECD (2019), Digital Government Review of Argentina: Accelerating the Digitalisation of the Public Sector, OECD Publishing, Paris, https://doi.org/10.1787/354732cc-en

[78] E-estonia (2018), X-Road, https://e-estonia.com/solutions/interoperability-services/x-road

[79] OECD (2018), Open Government Data Report: Enhancing Policy Maturity for Sustainable Impact, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264305847-en

[80] Law No. 13,460 of 26 June 2017, available at: www.planalto.gov.br/ccivil_03/_ato2015-2018/2017/Lei/L13460.html

[81] CNIL, “What you should know about our standard on data protection governance”, https://www.cnil.fr/en/what-you-should-know-about-our-standard-data-protection-governance

[82] Singapore, “Trusted Data Sharing Framework”, https://www.imda.gov.sg/-/media/Imda/Files/Programme/AI-Data-Innovation/Trusted-Data-Sharing-Framework.pdf

[83] XRBL Introduction, https://www.xbrl.org/the-standard/what/an-introduction-to-xbrl/

[84] The Open Contracting Data Standard, https://www.open-contracting.org/data-standard/