Abstract
National Statistical Offices are increasingly expected to expand their role beyond traditional statistical production to act as central stewards of national data ecosystems. This expansion creates a fundamental institutional tension: organisations established to guarantee professional independence and confidentiality are now being asked to promote data sharing, reuse, and policy analytics. This article examines Lithuania as a critical case of this transformation. In 2023, Statistics Lithuania was legally reconstituted as the State Data Agency (SDA), receiving a mandate that integrates official statistics with nationwide data governance. Drawing on legal analysis, institutional documentation, and operational evidence, the study explores how Lithuania has centralised administrative data flows, established a national data lake, and developed secure analytical environments to support ministries and municipalities. Concrete applications in health surveillance, social policy, and municipal planning demonstrate the benefits of this model, including reduced respondent burden, more timely indicators, and enhanced evidence-based decision-making. At the same time, the Lithuanian experience reveals significant governance risks related to confidentiality, politicisation, and technological dependence. This paper argues that NSO-led data agencies can strengthen public sector capacity only if robust legal safeguards, transparent access procedures, and ethical oversight evolve alongside technological systems. Lithuania thus provides both inspiration and caution for countries seeking to reposition their statistical authorities at the centre of the digital state.
Introduction
Declining survey response rates, the rapid growth of administrative and digital data, and increasing demand for real-time policy evidence are transforming the role of official statistics, see, for example.1,2 Governments now expect statistical authorities not only to produce indicators but also to manage data infrastructures, integrate multiple sources, and facilitate access to microdata for research and policymaking. This evolution has led to the emergence of NSOs as “trusted data intermediaries” within national data ecosystems.3,4,5,6
Yet this expansion creates a fundamental institutional dilemma. Official statistics are built on principles of professional independence, confidentiality, and methodological integrity. 7 Data agencies, by contrast, are designed to promote interoperability, data sharing, and policy analytics. When these two roles are merged within a single organization, tensions arise between protecting sensitive information and enabling broad reuse, between political neutrality and active policy support, and between statistical quality and the opportunistic use of administrative data.
Lithuania provides a particularly revealing case of this transformation. In 2023, Statistics Lithuania was legally reconstituted as the State Data Agency (SDA), receiving a mandate that goes far beyond conventional statistical production. The new institution not only continues to compile official statistics but also acts as the central hub for state data governance, operating a national data lake, integrating registers across government, and supporting ministries and municipalities with advanced analytical tools. All public authorities are now legally required to transmit administrative data to the SDA within thirty calendar days, marking a decisive shift from fragmented data ownership to centralized stewardship.
This article examines the Lithuanian experience as a test case for NSO-led data governance. It asks three main questions. First, how can a national statistical office be transformed into a state data agency while preserving statistical independence and confidentiality? Second, what legal, institutional, and technological mechanisms enable large-scale reuse of administrative data without undermining public trust? Third, what governance risks may emerge when statistical authorities become central actors in policy analytics?
The study adopts a qualitative institutional analysis based on legal texts, policy documents, and operational evidence from the Lithuanian State Data Agency, complemented by comparative insights from other European models. Rather than treating Lithuania simply as an example of innovation, the paper uses it as a critical case to explore the broader question of whether NSO-led data agencies can reconcile the competing imperatives of statistical integrity and data-driven governance.
The Lithuanian model offers significant advantages: faster data integration, reduced respondent burden, richer evidence for policymaking, and new capacities at municipal level. At the same time, it raises important concerns regarding confidentiality, politicization, vendor dependence, and the concentration of informational power. Understanding these trade-offs is essential not only for Lithuania but for all countries where statistical systems are being asked to operate at the centre of national data ecosystems.
The article proceeds as follows. Section 2 outlines the legal and institutional transformation that created the Lithuanian State Data Agency. Section 3 examines the expansion of administrative data sources and the development of the national data infrastructure. Section 4 presents concrete applications of this infrastructure in health, social policy, and municipal governance. Section 5 analyses the governance risks and challenges associated with the model. The final Section 6 discusses the implications of the Lithuanian experience for the future of official statistics in the digital state.
The new Lithuania's legislative framework
The State Data Agency and Statistics Lithuania have more than one hundred years of history (https://vda.lrv.lt/en/news/lithuanian-statistics-turns-106-today-X2x/). During the Soviet era, the Lithuanian economy was centrally planned, and Statistics Lithuania focused on industrial production and expansion in line with centralised Soviet directives. Following the Soviet period, Statistics Lithuania supported the development of the country's new civil registration system.
The legal framework underpinning these changes has evolved incrementally yet decisively over the past few years. In October 2020, the legal basis for official statistics was revised with limited amendments. In July 2021, Lithuania adopted the Law on the Right to Obtain Information and Reuse of Data, establishing the foundation for broader data sharing beyond statistical purposes. This was followed by the Health Data Reuse Act in July 2022, which introduced specific provisions for one of the most sensitive and valuable categories of state data. During the COVID-19 pandemic, the government placed its trust in Statistics Lithuania not only because of its established reputation but also because Statistics Lithuania was already using Palantir technology to support the rapid analysis of health data.
A major milestone was reached in 2023, when the entirely new Law on Official Statistics and State Data Governance (https://www.e-tar.lt/portal/lt/legalAct/TAR.026F44E06A27/asr) came into force, formalizing the establishment of the State Data Agency and embedding governance functions directly into law. Finally, the adoption of the EU Data Governance Act in September 2023 marked another significant step forward.8,9
The State Data Agency (SDA) now performs several key functions that extend well beyond traditional statistical production. While it remains the official producer of national statistics, it has also been entrusted with providing statistical data for scientific research, thereby facilitating evidence-based inquiry. As a central hub for state data reuse, the SDA operates technology platforms to prepare data for open access, ensures confidentiality, and assesses the readiness of data for publication.
Under the current legal framework, institutions are required to provide the State Data Agency, free of charge and within a maximum of thirty calendar days, with copies of primary and administrative data, as well as any related information or documentation necessary for the Agency's effective performance of its functions. Public administrations, like those in other sectors, are often reluctant to share data with the National Statistical Office (NSO) or other entities, even under strict confidentiality protocols. In Lithuania, following the enactment of the new Data Law in 2023, administrations are required to share data with the Lithuanian State Data Agency within 30 calendar days. Figure 1 presents the timeline of key legislative developments. A shift in attitude is evident: data sharing has become widespread and extensive, partly because the government has benefited from improved policy implementation. Technical data governance 10 reforms address broader democratic governance objectives, 11 creating a bridge between administrative modernization and civic legitimacy.
Timeline of key legislative milestones shaping Lithuania's transition from a traditional National Statistical Office to a State Data Agency with responsibilities for official statistics and state data governance.
This system will be managed in accordance with the principles of professional independence, objectivity, quality, confidentiality, integrity, accessibility, lawfulness, accuracy, and semantic compatibility. The law defines the rights and responsibilities of both data providers and data users to ensure the responsible and appropriate handling of data. Accordingly, the legal framework establishes obligations regarding data accuracy and security. The SDA must develop an annual State Data Governance Programme, specifying activity names, responsible agencies, legal bases, data sources, and funding sources. All data management must comply with Regulation (EU) 2016/679 (GDPR), prioritizing data protection. Official statistics sources include data from respondents, administrative sources, and other legal entities. The law is designed to align with EU regulations on data governance, confidentiality, and protection. 9
Traditionally, administrative offices in Lithuania, as in most European countries, treated data as institutional property, held and used primarily for internal administrative purposes. Data sharing between agencies was often fragmented and bureaucratically complex. The new law marks a paradigm shift by viewing data as a public asset rather than the exclusive resource of an individual agency.
Is it possible that a new law in the coming years could reduce the level of confidentiality? A law that prioritizes data reuse and sharing might, in practice, weaken confidentiality protections hypothetically if institutional capacity, oversight, and ethical safeguards do not evolve at the same pace. The challenge for Lithuania, as for other countries, will be to expand access without eroding trust—ensuring that transparency and confidentiality remain mutually reinforcing rather than contradictory.
The State Data Agency (SDA) of Lithuania has carried out an extensive inventory of state data sources (The data of this section are updated to January 2026: https://duomenys.stat.gov.lt/). The SDA now has access to a vast range of new data sources, but not all of these data constitute ‘statistical data’. In principle, all administrative data can be included in this inventory. This mapping effort identified 1935 data sources, of which 480 have been made available for wider use. These are supported by 1219 metadata descriptions, ensuring discoverability and standardization, as well as hundreds of institutional connections (779) that link datasets into coherent flows. In total, over 550 institutions contribute to this ecosystem. Such efforts align with global trends in open data inventories and dashboards, enhancing data availability and usability.
Attention has been devoted to the reuse of health data (from COVID-19), which is both sensitive and highly valuable. The State Data Agency has established procedures for handling applications for health data access, ensuring that reuse complies with legislation and ethical standards. Demand for health data has grown steadily, with applications increasing between 2023 and 2025. In 2023, there were 10 applications on reusing health data after several revisions and corrections, while in 2024, the number doubled to 20 applications. This trend reflects the rising importance of health data in research, governance, and innovation, consistent with broader European debates on data protection and innovation.9,12 Lithuania has institutionalized a comprehensive data governance program that integrates legislation, technology, and organizational practices. The Law on Official Statistics and State Data Governance mandates that all users of state data must be legally grounded and registered within the governance program. The technological foundation of this governance model is a state data platform that consolidates fresh data from multiple sources and provides a secure analytical environment equipped with a multifunctional programming environment, such as Python, R, PySpark, SQL, GIT, AI tools, and business intelligence (BI) tools, which are all supported by Palantir's technology.
The SDA has institutionalized the provision of data to public institutions, enabling ministries, agencies, and municipalities to rely on secure and timely data flows. More than 5000 users have registered, with over 2000 active users accessing the system in the past three months. The Ministry of Social Security and Labor, Ministry of Environment, National Audit Office, and Ministry of Finance are among the most frequent users. At the core of this system is the State Data Lake, which consolidates and organizes data with a capacity exceeding 120 terabytes and a processing capacity of 100,000 CPU hours per month. This approach aligns with international trends in developing centralized yet flexible state data infrastructures. 13 There were more than 12,000 microdata tables that were produced.
The technological transformation has been driven by the deployment of the Palantir Foundry platform, which supports both business intelligence (BI) tools for non-technical users and advanced programming environments for specialists. Since 2023, onboarding programmes have provided tailored guidance to each institution. In 2024, BI training reached more than 1500 participants, and in 2025, advanced training sessions further expanded analytical capabilities. These initiatives underscore the importance of developing human capacity alongside technical systems. For example, in 2025, many training activities were conducted by SDA to train the administration in using the data platform and to improve data literacy in schools. Training sessions involved municipal staff, ministry employees, teachers, and municipal enterprises.
All new data sources for the Data Lake are funded by the Recovery and Resilience Facility (RRF). The unique geopolitical situation in Europe and Lithuania necessitates enhancing the state's resilience to threats. The project, “Integrating State Information Resources into a Data Lake’, is funded by the Recovery and Resilience Facility (RRF). Its aim is to catalogue all Lithuanian public sector data, integrate and centralise it, and make it accessible, thereby strengthening the state's resilience to threats. In this geopolitical context, Lithuania's data represent a valuable strategic asset, making data storage a critical decision. The ability to transfer data to the cloud and the choice of storage location have become essential considerations. All data are stored in a NATO member country.
Lithuania's state data infrastructure in practice
Applications already demonstrate the impact of this new infrastructure: integrating government investment data, monitoring the National Development Plan and Sustainable Development Goals (https://sdg.stat.gov.lt/) (SDGs), assessing the effectiveness of legislation, and conducting budgetary impact assessments. State data has also been mobilized for crisis-related applications, such as monitoring refugee flows from Ukraine and restricted goods exports. At the municipal level, the DataLab provides education, civil protection, and infrastructure modules, illustrating how local authorities can benefit from national data governance reforms. For those new tasks, by 2023, Statistics Lithuania had hired more than 100 employees, bringing its current workforce to approximately 600.
Following two health applications, in May 2025, the Lithuanian State Data Agency (SDA), in collaboration with the National Public Health Centre (NVSC), developed a new interactive Human Papillomavirus (HPV) vaccination monitoring dashboard. The tool integrates data from multiple administrative sources — including population registries, health service records, and primary care assignments — to enable detailed tracking of HPV vaccination status across age groups, sexes, municipalities, and healthcare providers. The dashboard addresses longstanding issues with fragmented and inconsistent vaccination records by aligning records across systems and adjusting for population mobility and varied data reporting standards. Users can visualize vaccination coverage at regional and local levels, filter by demographic groups, and observe trends over time. This initiative demonstrates how SDA's integrated data governance infrastructure can support nuanced public health surveillance while ensuring data privacy through anonymized records.
In November 2025, SDA published results from a new statistical study on household expenses in pharmacies and pharmacy goods sales. For the first time, this analysis quantifies how much Lithuanian residents spend on pharmaceutical products, distinguishing between non-prescription medicines, non-compensated prescription medicines, dietary supplements, and state-compensated medicines. Using integrated data sources — including pharmacy receipt records, e-prescription systems, and regulatory registers — the study found that in September 2025 households spent over
In June 2023, an analysis of SDA debt conducted for Lithuania's Ministry of Social Security and Labor (SADM) revealed that approximately 70% of working-age individuals with outstanding debts are either unemployed or belong to the lowest income bracket. Following the publication of these findings, SADM revised its policies (https://socmin.lrv.lt/lt/naujienos/4-is-5-darbingo-amziaus-skolininku-nedirba-sadm-siulymas-diferencijuotas-skolu-isieskojimas/). This challenge has prompted the ministry to propose a differentiated debt recovery model aimed at reducing social exclusion and encouraging formal employment.
After analysing the data from the State Data Agency (SDA), the ministry's suggested reform suggests lowering wage garnishment rates for individuals earning minimum wage to encourage debtors to enter the formal labour market. All data and statistics were managed by the SDA. There are several examples of data impact policies shaping legislation across various sectors.
While the Lithuanian State Data Agency's primary mandate is to produce economic and social statistics, its capabilities can be expanded and integrated into national defence strategies, particularly in a complex geopolitical context. A variety of applications were implemented between SDA and the administrations. Refugees flow from Ukraine received special attention. A change of law was made because an analysis of data of vehicle revision highlighted fraud. Many vehicles in the revision registered less kilometres. Another issue was when, after a comprehensive data analysis was done among police registry, social insurance registry, civil registry, real estate registry and migration registry, the change of compensating home renovations was implemented.
At municipality level, a special attention is dedicated to infrastructure, education and civil protection. The Lithuanian State Data Agency's Municipal DataLab provides municipalities with a unified, register-based analytical environment that supports infrastructure planning, civil protection, and education policy using highly granular administrative data.
In the infrastructure module, municipalities can observe every real-estate object within their territory down to street and building level, including construction year, energy performance category, and renovation status. These linked datasets allow municipal infrastructure departments to identify clusters of poorly insulated buildings, monitor renovation progress, and prioritize investments based on objective indicators rather than fragmented local records. This enables more strategic planning of energy efficiency programmes and capital works at a level of detail that would be impossible using conventional surveys or reports.
The civil protection module integrates population registers, disability records, and shelter infrastructure to support emergency preparedness. Municipalities can assess walking distance to the nearest collective protection structure, evaluate shelter capacity relative to the spatial distribution of seniors and persons with disabilities, and simulate evacuation needs for vulnerable populations. This transforms civil protection from a static compliance exercise into a continuously updated, data-driven planning function.
The education module links school registers and enrolment flows to track student mobility between schools, class sizes, and absenteeism patterns. Education departments can monitor inflows and outflows by grade and institution, identify emerging capacity pressures, and detect early warning signals such as rising non-attendance in particular subjects or schools. This supports more equitable class formation, targeted interventions, and more precise allocation of teaching resources. Together, these three applications illustrate how the SDA's integrated data infrastructure enables municipalities to move from fragmented administrative reporting to real-time, evidence-based governance across core public services.
Potential issues and challenges
Recasting a national statistical office as a data agency could be the most effective way to harness administrative data for official statistical purposes instead of a fragmented system. However, its success depends on effective governance and a robust legal framework. The Lithuanian model of state data governance exemplifies ambition, innovation, and strong data leadership by its national statistics office. Nevertheless, it also faces significant challenges and risks. A critical perspective reveals at least three dimensions where the reforms encounter potential issues: legal and ethical dilemmas, institutional and political tensions, and technical and operational constraints.
Firstly, the expansion of state data governance raises complex questions regarding privacy, consent, and the scope of state authority. The reuse of health data, for example, is often promoted as a driver of innovation; however, critics argue that it risks normalizing secondary uses of sensitive data without meaningful citizen consent. Although legal safeguards exist, the rapid pace of emergency data sharing (within five working days) may circumvent adequate ethical review. Moreover, the ambiguity between ‘statistical data’ and ‘state data’ risks creating legal loopholes that could undermine both transparency 14 and accountability. Lithuania is not unique in this regard; this highlights the persistent tension between open data and data protection. In Lithuania, open data is the responsibility of the government and the SDA, creating a potential conflict of interest. Reister 15 further addresses the issue of National Statistical Offices (NSOs) assuming data stewardship roles, cautioning against merging these responsibilities with the production of official statistics. The inherent conflicts, such as the tension between data confidentiality in official statistics and the promotion of broader data use, may undermine the integrity of traditional statistical practices. The article emphasizes the necessity of carefully evaluating the fitness for purpose of data, distinguishing between what is appropriate for statistical aggregation and what is suitable for individual-level access, particularly in the use of administrative data (see also 16 ). Lithuania should consider establishing a comprehensive data governance framework to address these challenges when utilizing sources beyond official statistics. A new law that prioritizes data reuse and sharing can, in practice, diminish confidentiality protections if institutional capacity, oversight, and ethical safeguards do not advance at the same pace. The challenge for Lithuania, as for other countries, will be to expand access without eroding trust — ensuring that transparency and confidentiality remain mutually reinforcing rather than contradictory. This is because the SDA is dedicated to transparency and to explaining the new tasks.
Secondly, from an institutional perspective, the centralization of authority within the State Data Agency (SDA) can create both efficiencies and vulnerabilities. While consolidation enhances coordination, it also risks concentrating excessive power in a single entity, potentially leading to the politicization of data flows. If the SDA becomes closely aligned with the priorities of the current government, data selection, release, and interpretation could reflect political biases rather than neutral evidence. 11 Additionally, ministries and municipalities may feel disempowered by the strong central hub, which could discourage local innovation or reinforce hierarchical control over data access. From another perspective, this can be seen as a decidedly empowering step: municipalities or ministries will no longer need to worry about the technical aspects of data collection or access and will be able to focus entirely on the decision-making process. Capano et al. 17 note that political resistance or bias from politicians can hinder data-driven decision-making. They also address the frequent ambiguity surrounding the role of individual agents within policy process theories. The authors propose a new framework using Merton's concept of function to analyse agency by identifying four main functions—steering, innovation, intermediation, and intelligence—and their associated patterns of action. These functions can be enacted by various agents in different ways, influenced by individual motivations and structural factors within the policy subsystem. This framework provides a more nuanced understanding of how agents influence policy dynamics, moving beyond simplified models.
Finally, the technological infrastructure—particularly the reliance on a single commercial platform such as Palantir Foundry—raises concerns about vendor lock-in, transparency, and long-term sustainability. While powerful, proprietary systems can be costly, difficult to customize, and dependent on external providers. This dependence could undermine the sovereignty of Lithuania's data governance model, creating vulnerabilities linked to multinational technology firms. Furthermore, despite extensive training efforts, a significant capacity gap persists between highly skilled analysts and ordinary civil servants. Without sustained investment in human resources, there is a risk that only a small elite within the government will be able to fully leverage these tools, thereby limiting the broader democratization of data use. There is an ongoing project funded by the government to replace the technology of Palantir. This funding will hopefully bridge the existing ‘capacity gap’. At the same time, we must acknowledge that similar contracts between government agencies and major technology companies exist at various levels in countries worldwide.
Taken together, these three dimensions suggest that the Lithuanian model is innovative yet potentially challenging. Legal safeguards may not fully protect against ethical risks; institutional centralization could inadvertently foster politicization; and technical reliance on proprietary platforms raises concerns about long-term sustainability. If left unaddressed, these issues could erode trust in the system, undermine its legitimacy, and compromise its democratic promise. 11
Conclusions
By transforming Statistics Lithuania into the State Data Agency, embedding governance within the law, and developing robust technological infrastructure, the country has moved beyond traditional statistical production 18 to adopt a comprehensive model of state data stewardship. This model integrates legislation, technology, and human capital into a unified system designed to promote transparency, accountability, and evidence-based governance.
Lithuania is gaining prominence in the field of official statistics. In May 2026, Lithuania will host the 20th IAOS Conference in Vilnius, organised jointly by the International Association for Official Statistics (IAOS) and the State Data Agency (Statistics Lithuania). The three-day event, themed “Navigating the Data Revolution: Innovations and Impact in Modern Statistics”, will bring together statisticians and data professionals from around the world to discuss innovation, ethical challenges, and the future role of official statistics in an era of data abundance and technological change. International recognition of the Agency was demonstrated by the election of Dr Jūratė Petrauskienė, Director of the SDA, as President-Elect of the IAOS in April 2025. In this role, she will assume the association's presidency beginning in 2027, highlighting Lithuania's growing influence within the global official statistics community and reinforcing the country's leadership in data governance and statistical innovation.
Lithuania has also secured substantial support from the European Union to further develop its data and innovation ecosystem. In 2025, the country was awarded
Complementing these institutional achievements, Lithuania has also made significant progress in open data maturity. In the European 2025 Open Data Maturity assessment (https://data.europa.eu/en/open-data-maturity/2025), Lithuania ranked second in Europe, achieving high scores across policy, portals, data quality, and impact dimensions. This performance reflects the SDA's strategic efforts to make data reliable, accessible, and usable for society, public sector innovation, and evidence-based decision-making — further demonstrating how integrated data governance can strengthen national statistical capacity.
In conclusion, the Lithuanian reform demonstrates how transforming a National Statistical Office into a broader data authority can unlock major public value, but only if governance evolves as quickly as technology. The experience of Lithuania shows that legal clarity, ethical oversight, and institutional balance are as crucial as technical innovation. Safeguarding confidentiality while enabling reuse, preventing over-centralization within the State Data Agency, and reducing dependence on platforms like Palantir will determine long-term legitimacy. Ultimately, trust—not infrastructure alone—will decide whether data governance strengthens democracy or quietly restrains it. The long-term outcome will depend on how these tensions are managed.
Footnotes
Acknowledgments
Acknowledges funding from Horizon Europe through the Marie Skłodowska-Curie Actions. Thanks are due to Jūratė Petrauskienė and the State Data Agency of Lithuania (Valstybės duomenų agentūra) for their hospitality and support. I am especially grateful to Vadimas Ivanovas for his presentations and to Jonas Bačelis for his exceptional support. I had great and inspiring meetings with Daiva Jurelevičienė, Gita Literskė, Tomas Rudys, and Inga Masiulaitytė-Šukevič. I am also grateful for the insightful meetings with the Members of the Lithuanian Parliament, Mr Lukas Saviškas and Ms Giedrė Balčytytė. Thank you for the inspiring meeting, Rimantas Žylius, Palantir.
Funding
The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by Horizon Europe: Marie Sklodowska-Curie Actions (Project ID: 101105704).
Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
