The evolving role of Official Statistics in the digital era—the case of Switzerland

1 Introduction

Data is increasingly recognized as a critical resource for prosperity, sustainable economies, scientific progress and effective policymaking.^1–3 Unlocking this potential depends on the responsible and efficient use of data that is high-quality, accessible, timely, and reliable - qualities that are indispensable for public sector innovation, for official statistics, and for evidence-based decision-making in general. More broadly, well-structured data ecosystems, clear governance frameworks and the capacity to create value from data are central to modernizing societies.

However, despite the ubiquity of data and digital technologies, data literacy remains limited across the public sector. ⁴ Debates on digital transformation are frequently dominated by technology-oriented narratives (e.g., cloud computing or AI). This often postpones efforts to strengthen core data foundations for these technologies, such as (meta-) data documentation (e.g., findability) and interoperability. Developing data literacy - the capacity to interpret, contextualize and govern data in ways that generate societal value - has therefore become a core requirement for public administration.

Without a focused approach to making data visible and interoperable, any efforts to derive additional value from data are likely to be more difficult and costly. Addressing these challenges requires a common language and establishing collaborative data governance frameworks. While definitions of “data literacy” may vary,^5,6 the capacity to address these issues is probably common to most.

In this setting, statistical offices hold a unique advantage. They combine long-standing methodological expertise with robust quality frameworks and a democratically legitimized mandate. They possess a high degree of data literacy, particularly in translating between technical and non-technical questions of data use and re-use. Together, these skills constitute a strong foundation for an enhanced role in the (digital) transformation of public administration.

As an example, the Swiss Federal Statistical Office (FSO) has emerged as a central actor in strengthening data governance and data literacy within the Swiss federal administration. Since 2019, successive decisions have expanded its responsibilities to include interoperability, stewardship, analytical capacities and innovative methods. These developments extend the FSO's role beyond official statistics and position it as a key player in the federal government's digital transformation.

This article uses the FSOs trajectory as a case study to document how the role of a National Statistical Office has evolved over time in response to legal mandates, political expectations, and organizational dynamics within a democratic data ecosystem. It describes observed developments, institutional arrangements, and practical experiences, and reflects on both the enabling effects and the challenges, tensions, and unresolved issues associated with the expansion into data management and data science.

Taken together, the developments described reflect an incremental and learning-based evolution, shaped by both proactive institutional positioning within the FSO and reactive adaptation to political mandates. The following chapters first present the specific mandates and their background and then provide examples of the FSO's contributions to digital transformation, highlighting key learnings and their connection to its statistical mandate.

2 Legal context: The federal statistics act as an enabler

To better understand the recent advances of the FSO mentioned above, it is instructive to look at the Federal Statistics Act of 1992, ⁷ which, despite its age, still underpins many of the developments described in this article. It defines official statistics as an essential information infrastructure for a modern democratic state and establishes two fundamental principles for statistical production that have shaped the development of official statistics in Switzerland ever since.

First, it mandates an output-oriented design of statistical products, based on the information needs of users while specific input data and sources are specified only in broad terms. The principle is reinforced by legal provisions requiring the formal involvement of user groups from the public sector, business and research, thus institutionalizing a multi-stakeholder approach.

Second, the rules for data collection outlined in article 4 of the Act establish a clear hierarchy of data sources. Surveys may not be conducted if the necessary data already exists at the federal administrative level or are generated in the implementation of federal legislation. Additional data should first be obtained from cantonal or communal authorities or other public bodies. Direct surveys are a last resort, limited to what is strictly necessary.

Following the adoption of the Federal Statistics Act in 1992, the FSO began a gradual transition by creating central registers (e.g., businesses, buildings and housing) and by improving access to population registers operated by the different cantons of Switzerland. This culminated in the introduction of a register-based population census system in 2010.

These legal provisions not only shaped the evolution of statistical production but also compelled the FSO to work in close coordination with federal and cantonal administrations. By relying on administrative data sources, the FSO developed practical expertise in intergovernmental collaboration, standardization and data integration. Acquired in the service of its statistical mandate, this expertise prepared the FSO to take on the additional mandates presented in the following chapters.

While the principles underpinning data reuse, coordination, and interoperability were thus established early on, their application initially remained largely confined to the domain of statistical production. The extension of these principles to the broader administrative context only gained momentum later, as political expectations increasingly called for more systematic data driven innovation across government.

3 Political dynamics: From census success to new mandates

Building on the legal foundations, the FSO experienced a period of calm after the above-mentioned introduction of the register-based census in 2010. However, the census's success in reducing respondent burden for the Swiss population soon triggered political debate, as Parliament demanded similar relief for businesses.

Initially in 2014, proposals were made to drastically reduce the FSO's capacity to conduct business surveys. ⁸ But, Major business organizations opposed these proposals, emphasizing their potential implications for Switzerland's economic policy foundations, and Parliament ultimately rejected them. The political debate then shifted towards strengthening the FSO's coordinating role and expanding the use of administrative data in the Swiss statistical system.

This shift culminated in a 2016 parliamentary motion ⁹ calling for a streamlined process to prevent multiple data requests to businesses. The Federal Council subsequently mandated the FSO to implement the necessary measures. The challenge was that this required the coordination of data collection for both administrative and statistical purposes and across multiple levels of government, without clarifying the FSO's authority in relation to other administrative units and agencies.

By then, the FSO had already begun to reflect proactively on its future role. Guided by the Federal Statistics Act, these internal considerations focused on expanding the use of administrative data, strengthening interoperability, and adopting new methods. Priorities included mapping available data and developing a comprehensive data catalogue to identify further sources. Cooperation from other units was seen as contingent on their deriving tangible benefits beyond official statistics. This dual dynamic - external political pressure and internal foresight - shaped the subsequent mandates, designed both to add value for other agencies and to strengthen statistical production.

A symbolic milestone came in 2018, when the Federal Council visited the FSO for one of its rare ‘extra muros’ retreats and dedicated it to the Office's role in the national data ecosystem. As Switzerland's seven-member executive government acting collectively as both government and head of state, the Federal Council plays a central role in defining strategic orientations for federal data policy. This marked the beginning of the FSO's transformation and following the retreat, the Federal Council gradually expanded the FSO's role.

This took the form of new mandates that formalized the Office's coordinating and data-governance functions. In 2019, the FSO was mandated to implement the “National Data Management” program, aligned with the Tallinn Declaration's “once only” principle ¹⁰ The program combines a technical infrastructure for metadata documentation (a national data catalogue: I14Y described below) with data stewardship to harmonize data across federal and cantonal administrations. In addition, the Federal Council's decisions to establish the Data Science Competence Centre (DSCC) ¹¹ in 2020 and the Competence Network for Artificial Intelligence (CNAI) ¹² in 2022, both located at the FSO, were taken to foster the skills needed to further profit from data through innovative methods.

Building on the political developments outlined above, the current role of the FSO is defined by three complementary mandates: (a) “statistics,” (b) “data management,” and (c) “data science and artificial intelligence.” Taken together, these mandates illustrate the FSO's dual function: safeguarding its traditional statistical tasks, while taking on broader coordination and enabling responsibilities.

4 Data management mandate: Enabling interoperability and reuse

The FSO's ‘data management’ mandate emphasizes its enabling role within the federal administration. This mandate covers a broad set of tasks and includes data stewardship roles as well as data governance. To accomplish these tasks, the FSO adopts a socio-technical approach, which views technical infrastructures and social structures as interdependent and designs them jointly. In practice, this means combining infrastructures such as the I14Y Interoperability Platform described below, which makes data visible, with governance roles defined through data stewardship - which build on this visibility.

Data stewardship is central to this mandate. Stewards facilitate harmonization, standardization and foster trust between administrative units. Under the National Data Management (NaDB) program launched in 2019, ¹³ the FSO introduced a dual role model: Data Owners are responsible for specific datasets, while domain-specific Data Stewards ensure harmonization and compliance with interoperability guidelines. At the strategic level, the Swiss Data Steward (formally the Director General of the FSO) coordinates these activities across thematic domains and government levels. The model aligns with international frameworks,^14,15 though a detailed discussion would go beyond the scope of this article.

An important step in consolidating and clarifying the FSO's mandate is the adoption of the Federal Act on the Use of Electronic Means for the Fulfilment of Government Tasks (EMBAG) ¹⁶ in 2024 and, subsequently, the Digitalization Ordinance (DigiV) ¹⁷ in 2025. Especially with the DigiV, key principles regarding data management are operationalized and aligned across the federal administration. For the FSO, this results in a clearer codification of its tasks that had previously evolved through practice and political mandates, particularly in the areas of data stewardship, coordination, and data governance.

On the technical side, the I14Y Interoperability Platform, ¹⁸ launched in 2021, serves as Switzerland's national metadata catalogue. On I14Y, authorities across all federal levels can describe their operational datasets, (around 1600 by the end of 2025) associated electronic interfaces (APIs, around 70 by the end of 2025), and digital administrative services (around 125 by the end of 2025) through metadata, thereby making them findable and accessible. A global search function allows users to discover all objects within the catalogue, supporting transparency, interoperability, and reuse.

In parallel, the FSO participates in the implementation of the federal master data management strategy, which aims to gradually harmonize master data in foundational domains such as businesses, persons, spatial data, and buildings. These efforts led by the Federal Chancellery's Digital Transformation and ICT Steering Sector (DTI) ensure that essential master data are maintained consistently, legally compliant, and reusable across the whole of the federal administration. ¹⁹

Finally, the FSO is the leading agency for Open Government Data (OGD) in Switzerland. It acts as the secretariat of the federal OGD strategy and is Switzerland's largest provider of open data. Through opendata.swiss ²⁰ a platform operated by the FSO, administrative agencies from all levels of Government as well as other institutions systematically publish datasets, enhancing transparency and reuse by government, academia, business, and civil society.

5 Data science and AI mandate: Fostering innovation and trust

Under the “data science and AI mandate,” the FSO acts as a facilitator. It provides tools, expertise, and frameworks that enable government agencies to apply advanced methods and generate insights from data. Within the FSO, data science primarily refers to analytical methods supporting data integration, quality assurance, and exploration, whereas artificial intelligence encompasses a narrower set of techniques involving higher degrees of automation.

As a key component the FSO's Data Science Competence Centre (DSCC) plays a central role in embedding data science methods in the federal administration and fostering a culture of collaboration across different levels of government. At the heart of this function is the DSCC's suite of data science services for government (“data science as a service”), which includes project support, consultancy, and the dissemination of state-of-the-art data science knowledge. The DSCC carries out a wide range of projects for partner organizations in the federal administration—from building or optimizing data pipelines to improving the analytical use of agencies’ own data.

One example is a project to improve the quality of traffic data, where the DSCC works with the Federal Roads Office (ASTRA) to automate the detection of sensor errors on national highways and reconstruct missing values. Another example is LOMAS, ²¹ an open-source platform that enables authorized researchers and analysts to remotely execute analytical algorithms on sensitive datasets without accessing the underlying raw data. This approach has the potential to transform previously untapped administrative data into actionable insights, supporting innovation without compromising privacy. To provide the technical platform for LOMAS, the FSO is also developing dedicated Microdata Center, which will additionally offer secure access to (linked and pseudonymized) microdata for research institutions.

The FSO's commitment to a more data-driven administration is also reflected in its leading contribution to the Federal Strategy for Data Science (2022), ²² which promotes the coordinated use of data science across the administration by strengthening awareness, accessibility and collaboration. Furthermore, the FSO played a leading role in drafting the Federal Code of Conduct for Human-Centered and Trustworthy Data Science (2023). ²³ This code sets out ethical and professional guidelines for responsible use of data science in the federal administration, addressing issues of privacy, transparency and accountability.

Finally, the FSO's Competence Network for Artificial Intelligence (CNAI) ²⁴ plays a coordinating role in harnessing the potential of AI for the public sector. Established in 2022 by the Federal Council, the CNAI acts as a central enabler and facilitator of AI adoption across the federal administration. It provides a shared terminology foundation, consolidates expertise in a structured knowledge database, and fosters exchange through both a community of practice and a curated network of experts. CNAI also publishes an overview of ongoing AI projects in the federal administration, ensuring transparency and helping build trust in AI applications across government.

6 Reaping the benefits for official statistics

Although the FSO's additional mandates place it at the forefront of developing the Swiss data ecosystem, a central objective is to ensure that the outputs of these mandates contribute to improving official statistics. The potential added value lies in broadening the range of data available, embedding new analytical methods, and fostering collaboration within and beyond the federal administration. In practice, this added value is gradually emerging through improved data coverage, a progressive reduction in reliance on primary data collection, and expanding possibilities to analyze complex social and economic phenomena using linked administrative sources.

The data management mandate has already produced tangible results, such as the establishment of stewardship roles, the interoperability platform (I14Y), and ongoing harmonization and standardization efforts. These initiatives expand the potential database for statistical production, though their impact depends on the further development of governance frameworks and the willingness of diverse actors to share and reuse data. Where such conditions are beginning to be met, administrative data collected for non-statistical purposes can increasingly be integrated into statistical production processes. In this context, the FSO's legal authority to link administrative data at the population level represents an important comparative advantage, creating opportunities for more comprehensive analyses and deeper insights.

The data science and AI mandate opens additional perspectives for enhancing official statistics. With the DSCC, first steps have been taken to more directly integrate data science and artificial intelligence into statistical production. It is clear that the potential lies in applying advanced analytical techniques to improve data preparation, integration, and analysis. While many applications are still experimental or in pilot phases, they illustrate how data science can be systematically embedded into statistical production.

The expansion of mandates inevitably raises ethical and legal challenges. Safeguarding the professional independence and credibility of official statistics—anchored in the Federal Statistics Act—requires ongoing dialogue and cooperation across all levels of Switzerland's federal system. At the same time, the FSO has taken internal measures to maintain trust: separating statistical from non-statistical tasks, ensuring strict distinctions between statistical and administrative data, strengthening transparent communication, and regularly reviewing the legal and ethical framework. In this way, the Office seeks to ensure that its new responsibilities reinforce, rather than undermine, the foundations of trustworthy statistics.

Together, the FSO's mandates in data management, data science, and OGD illustrate a strategic shift: from producing statistics in isolation towards building a data-centric administrative ecosystem that, in turn, can strengthen official statistics. While only parts of this potential have so far been realized, the combined mandates point to a trajectory in which statistical production becomes more comprehensive, efficient, and relevant - anchored in collaboration, interoperability, and methodological innovation.

8 Key learnings and challenges

Implementing the FSO's new mandates required a paradigm shift in how public administration approaches data. Administrative bodies are naturally cautious in handling and protecting information, which makes cultural change particularly demanding.

Experience at the FSO with the Data Management Mandate, notably through the National Data Management Program (NaDB), indicates how essential a solid legal and governance foundation is. Early implementation efforts, including the introduction of stewardship roles and the development of the interoperability platform I14Y, were initiated in a context where legal provisions and formal governance structures were still evolving, which initially left responsibilities diffuse and progress limited. Only with the subsequent adoption of the Federal Act on the Use of Electronic Means for the Fulfilment of Government Tasks (EMBAG) in 2024 and Digitalization Ordinance (DigiV) in 2025) was the governance framework progressively consolidated, enabling stronger coordination and more systematic harmonization in the future.

Another key learning emerging from the FSO's experience concerns the sequencing of stewardship and infrastructure. It was initially assumed at the FSO that “visibility before reuse” would apply—i.e., that making datasets and APIs visible through I14Y would naturally lead to reuse. In practice, however, the experience suggests that awareness, competencies, and shared understanding had to be developed first. Embedded data governance through stewardship roles both at the FSO and in partner organizations contributed to building data literacy, creating the cultural and organizational basis for understanding the value of metadata catalogues.

Only now is this awareness gradually translating into effective use of I14Y. In short, the Data Management Mandate demonstrates that stewardship, legal clarity, and infrastructure must evolve in the right order: governance and literacy first, then technical visibility and reuse.

In the field of data science and AI, the picture emerging from the FSO's experience is equally heterogeneous. The Data Science Competence Centre (DSCC) has successfully implemented a range of collaborative projects with federal agencies and continues to do so. At the same time, experience shows that moving from innovative pilots to scalable and sustainable applications remains challenging, particularly with regard to data quality assurance, operational integration, and long-term ownership.

Initiatives such as LOMAS, which enable privacy-preserving analysis of sensitive microdata, illustrate potential that has been identified through concrete use cases, but which has yet to be systematically leveraged at scale. From an empirical perspective, these projects highlight both the opportunities and the institutional constraints associated with applying advanced methods in administrative contexts.

Overall, the FSO's experience suggests that technical infrastructures (such as I14Y), frameworks (such as data stewardship), data governance and cultural change (capacity building and data literacy) need to evolve in parallel. Where these elements have been aligned in practice, progress in data-driven innovation has tended to be more sustainable. Where alignment was incomplete, limitations and frictions became visible, underscoring the importance of a stepwise development.

9 The journey continues: A look ahead to 2030

According to its corporate strategy, ²⁵ the FSO aims to become Switzerland's central hub for data and information by 2030, serving the public, government, business and research. This vision reflects the FSO's commitment to creating value from data through data management and data science.

FSO will continue to fulfill its role in data harmonization and standardization, with the long-term goal of building a comprehensive data space for official statistics. This space is intended to facilitate seamless exchange and reuse of data across statistical domains and agencies, ranging from sensitive microdata to aggregated open datasets.

Operationally, the FSO is preparing for more automated and scalable production processes, enabling faster access to data while ensuring compliance with data protection and confidentiality. As new methodologies and technologies are adopted, the workforce will need to assume more innovative roles in an environment oriented towards continuous skills development.

Looking beyond Switzerland, the future of statistical systems depends increasingly on international cooperation. Trusted and interoperable data must be understood as a global public good, essential for effective governance in the digital age. As geopolitical tensions and algorithmic disinformation intensify, National Statistical Offices (NSOs) must not only produce reliable statistics but also safeguard the conditions under which data remains trustworthy.

As part of this agenda, the FSO has initiated a feasibility study for a Trusted Data Observatory (TDO) that could be installed at the UN in Geneva. Conceived as a neutral, technically grounded data catalogue, the TDO initiative connects international efforts in data interoperability and governance. Its purpose would not be to centralize data, but to provide metadata-based transparency about what data exists, how it is produced, and under what conditions it can be trusted. In this way, the TDO could support global standards, increase the discoverability of high-quality public data, and strengthen democratic legitimacy.

10 Conclusion

Looking at the FSO's trajectory, and based on the experiences documented in this article, the Office's longstanding expertise in statistics and data management has positioned it to play a central role in Switzerland's evolving data ecosystem. Its additional mandates in data management and in data science and AI represent more than functional extensions: they are part of a strategic evolution that seeks to combine established statistical competences with new approaches to interoperability, transparency and secondary data use.

The experience of recent years shows that the added value of these mandates is twofold: they create new opportunities for the production of official statistics, while at the same time strengthening the broader data ecosystem of the federal administration. Platforms such as the interoperability catalogue I14Y, the introduction of stewardship roles, and the establishment of the DSCC and CNAI have laid first foundations, but much of their potential remains to be realized.

At the same time, the FSO's role as a part of the international community of official Statistics extends beyond the national level. Trusted and interoperable data must be regarded as a global public good, and statistical offices—together with international organizations—are uniquely placed to safeguard the conditions under which data remains trustworthy.