Better skills data for smarter financing of education and training

5. Barriers to building integrated skills data systems

Improving the quality, accessibility and use of data in education and skills policy is not only a technical task but an institutional and systemic challenge. Countries often face multiple, interconnected obstacles across institutions, legal frameworks, capacities and technical systems that affect how data are collected, shared and used for policymaking. To support a clearer and more operational understanding of these challenges, this paper introduces a simple framework that groups the key barriers into four categories: institutional, governance and financing barriers; human-capital and analytical-capacity gaps; legal and regulatory constraints; and technical and interoperability challenges. This framework is designed to help policymakers identify where the main bottlenecks lie, understand their implications for data use, and prioritise the levers required to strengthen integrated skills data systems. The following sections discuss each of these four barrier types in detail.

5.1. Institutional and governance barriers

5.1.1. Fragmentation and lack of co‑ordination

Despite the potential of data to strengthen decision making, institutional and governance barriers often limit its effective use. A recurrent challenge is the fragmentation of mandates and responsibilities across the skills system. Education and training involve multiple ministries (education, labour, finance), several levels of government (national, regional, local) and a wide range of stakeholders, including public institutions, private providers and employers. In many countries, no single authority co‑ordinates a unified data strategy, and responsibilities for collecting and managing information are dispersed across agencies. Labour ministries may manage unemployment and training data, education ministries oversee school and higher education records, and regional authorities collect information on vocational programmes, each operating in separate systems. This fragmentation creates duplication in some areas, gaps in others and, critically, prevents the development of a coherent picture of how skills are developed, used and updated across the life course. OECD analyses consistently point to weak co‑ordination across ministries, levels of government and stakeholders as a major barrier to effective skills policy implementation (OECD, 2019[4]).

Emerging tools, including artificial intelligence, can help map existing data assets, harmonise definitions and support shared analytical platforms across institutions. However, their effectiveness depends on clear governance arrangements and agreed standards; without these, technology risks reinforcing rather than reducing fragmentation.

5.1.2. Incentives, culture and financing constraints

Weak co‑ordination is often reinforced by institutional cultures and incentive structures. Many administrations treat their datasets as proprietary resources, with limited trust in how other bodies may use the information. This reduces willingness to share data and limits the role of evidence in decision making. When data-driven practices are not embedded, leaders may rely on precedent or political judgement, and analytical work receives limited attention. Developing a culture that values evidence requires clear leadership, predictable rules for data use and incentives that encourage collaboration across institutions.

Financing and sustainability issues further constrain progress. Many improvements to data systems begin as short-term projects funded through pilots or external grants. When these end, systems often lack stable resources for maintenance, upgrades or analytical capacity. Budgeting rules may prioritise visible new programmes over long-term data infrastructure, and efficiency gains identified through better data use may be clawed back, reducing incentives to improve. Without predictable funding and aligned incentives, the benefits of data investments are difficult to sustain.

5.1.3. Capacity, accountability and policy leadership

Institutional arrangements also influence how effectively data can support policy. Responsibilities for outcomes are often distributed across multiple ministries and levels of government, yet accountability frameworks do not reflect these interdependencies. When no institution is clearly responsible for results that depend on shared data, gaps emerge between information and action. For example, education authorities may not engage with employment data if labour market outcomes fall under another ministry, even though such information is essential for programme planning.

Addressing these challenges requires both structural and leadership-driven reforms. OECD analysis shows that strong governance arrangements and explicit co‑ordination mechanisms are often decisive in shifting systems away from siloed practices and toward more integrated, outcome‑oriented approaches (OECD, 2024[11]). Effective governance also requires transparency about how data are produced – including collection methods, definitions, response rates and deviations from international standards – and clearly assigned responsibilities across the data lifecycle, from defining legitimate purposes and validating access to quality assurance, metadata maintenance and accountability for data reuse. Several OECD countries have taken steps to address these issues by creating central units or observatories with a mandate to integrate data across agencies and report to high-level bodies (see Box 1). Estonia’s Education Information System (EHIS) links education, employment and population data under the oversight of the Ministry of Education and Research. Ireland’s Skills and Labour Market Research Unit (SLMRU), hosted by SOLAS, provides integrated analysis to the National Skills Council. Norway’s Statistics Norway uses linked administrative datasets to support cross-ministerial decision making and reports jointly to the Ministry of Education and the Ministry of Labour. Chile’s Integrated System of Higher Education Information (SIES) consolidates education and labour data to inform national skills and financing policies. Australia’s National Centre for Vocational Education Research (NCVER) acts as the national custodian of VET data, while Jobs and Skills Australia provides independent analysis of current and future skills and workforce needs. The effectiveness of such arrangements depends on their integration with line ministries and on clear mechanisms to ensure that analytical findings translate into operational decisions, rather than remaining disconnected from policy processes.

Box 1. Country example: Finland

Finland’s Service Centre for Continuous Learning and Employment (SECLE)

Finland’s Service Centre for Continuous Learning and Employment (SECLE) provides an example of how institutional arrangements can help overcome fragmentation in skills systems. The Centre operates under the joint oversight of the Ministry of Education and Culture and the Ministry of Economic Affairs and Employment, two key actors in skills policy. This dual governance structure reflects the cross-cutting nature of skills development and helps align policy objectives across education and labour market domains.

SECLE’s mandate focusses on strengthening the skills of the working-age population and improving the match between labour market needs and workforce competencies. A central element of its approach is the development of integrated data capabilities. The Centre is investing in a comprehensive data infrastructure that links information on education, training and employment outcomes, combining existing administrative datasets with new digital tools and cloud-based solutions.

By creating a shared evidence base across institutions, SECLE supports more co‑ordinated decision making and helps bridge gaps between policy design and labour market outcomes. As a relatively recent initiative, systematic evidence on SECLE’s measurable impact on labour market outcomes is still emerging. This approach illustrates how governance reforms – through joint mandates, shared accountability and integrated data systems – can reduce fragmentation and improve the strategic use of data in skills policy.

Source: The Service Centre for Continuous Learning and Employment (2026[12]), About Us, www.jotpa.fi/en/about-us; Solita (2026[13]), SECLE supports the employment market with data, www.solita.fi/work/secle‑supports-the‑employment-market-with-data/.

5.2. Human-capital and analytical-capacity gaps

5.2.1. Shortage of specialised data talent

Even the most advanced data systems have limited value without the human capacity to interpret and act on the information they generate. Many OECD countries face a shortage of data engineers, analysts and scientists in the education and skills domain (OECD, 2023[14]) (OECD, 2024[15]). These specialists are in high demand across all sectors, and public administrations often struggle to attract and retain them due to less competitive salaries, limited career progression, and more rigid human resource frameworks (OECD Observatory of Public Sector Innovation, n.d.[16]).

This talent gap is particularly evident in advanced analytics. In many ministries of education or labour departments, only a small number of statisticians or data officers are responsible for managing large, complex datasets. Limited in-house capacity restricts the ability to conduct deeper analysis, increasing reliance on external consultants. While outsourcing can fill immediate needs, it is costly and risks fragmenting knowledge and reducing continuity within administrations. Emerging tools, including artificial intelligence, can help alleviate some of these constraints by automating routine data processing, supporting data linkage and enabling non-specialists to generate basic analysis. However, they do not substitute for analytical expertise and may increase the need for skills related to interpreting results, assessing model limitations and ensuring responsible use.

Talent retention is an additional challenge. The cost of competitive remuneration for such profiles is substantial and must be factored into the overall investment case for data system development. Even when governments succeed in recruiting skilled data analysts, retaining them can be difficult if career prospects and remuneration are more competitive in the private sector. OECD analysis has underlined that the scarcity of digital talent, particularly in areas such as data science, remains a significant barrier to digital transformation and that public sector organisations need to better integrate these profiles and methods into their daily operations (OECD Observatory of Public Sector Innovation, n.d.[16]).

Finally, continuous training for existing staff is another area of concern. Many officials responsible for managing education and labour market data have backgrounds in traditional statistics or IT, but may not be familiar with more recent methods of data analytics or visualisation. Without systematic professional development, staff skills risk becoming outdated. Targeted training in the use of new data tools, or in the interpretation of advanced analyses such as value‑added models of school performance, could substantially increase effectiveness. However, in many administrations, training opportunities and budgets remain limited, constraining the capacity to modernise data practices. Box 2 provides examples of how countries are developing structured approaches to strengthen digital and data skills across the public sector. While these initiatives represent promising approaches, systematic evidence on their cost-effectiveness and measurable impact on public sector performance is still limited.

Box 2. Country examples: Korea, Lithuania and Canada

Korea: National Human Resources Development Institute (NHI)

The Korean NHI, established in 2016 under the Ministry of Personnel Management, provides training and learning opportunities for civil servants across government functions. Its programmes combine e‑learning through the Talent Development Platform with regular curriculum courses.

A distinctive feature of the digital platform is its ability to track learning history, behaviour, and job profiles, allowing for personalised training recommendations based on individual skill levels.

The training primarily focusses on ICT fundamentals, artificial intelligence, and big data technologies, with an emphasis on their applications in government.

Source: Ministry of Personnel Management (2023[17]), Talent Development Platform, www.learning.go.kr/intro/board/viewPlatformIntroduce.do; Burtscher, M., S. Piano and B. Welby (2024[18]), “Developing skills for digital government: A review of good practices across OECD governments”, https://doi.org/10.1787/f4dab2e9‑en.

Canada: Canada School of Public Service (CSPS) Digital Academy

The CSPS was founded in 2018 to help civil servants build the skills needed for Canada’s digital transformation. It provides a structured framework for developing key competencies and offers training through live digital courses, self-paced modules, and mandatory programmes. These are complemented by informal opportunities such as learning events and online resources, making training accessible to over 350 000 federal civil servants.

The Digital Academy covers ten areas: digital government, product management, AI, digital leadership, agile, data, emerging technologies, service design, cybersecurity, and cloud. In data-related topics, courses include data visualisation, navigating large databases, and using data exploration tools.

Housed within the government’s School of Public Service, the Academy has a clear institutional mandate to drive skills development across the public sector.

Source: Government of Canada (2023[19]), CSPS Digital Academy, https://doi.org/10.1787/f4dab2e9‑en; Burtscher, M., S. Piano and B. Welby (2024[18]), “Developing skills for digital government: A review of good practices across OECD governments”, https://doi.org/10.1787/f4dab2e9-en.

Lithuania: GovTech Lab

Launched in 2019 as part of Innovation Agency Lithuania, GovTech Lab fosters public sector innovation by equipping organisations with tools, methods, and resources to solve challenges using technology. Its main activities include:

• GovTech Challenges: Co‑ordinating structured co-creation programmes between the public sector and startups to develop pilot solutions.

• Innovation Skills: Offering workshops and study visits to enhance public sector innovation

• Ecosystem and Community Building: Running international networks and programmes

Source: GovTech Lab (2024[20]) What is GovTech Labs, https://govtechlab.lt/about/; OECD (2024[21]), Enabling Digital Innovation in Government: The OECD GovTech Policy Framework, https://doi.org/10.1787/a51eb9b2‑en.

5.2.2. Data literacy and organisational barriers

Beyond technical specialists, a broader gap in data literacy persists among leadership and staff in many education and labour administrations. Senior decision makers – such as ministers, directors, or school leaders – often have limited training in interpreting statistical evidence. Where top officials are not comfortable with data, they may be less inclined to request rigorous analysis or to rely on it in decision making, even when robust information is available. This can reinforce reliance on intuition or political considerations. Building a culture of data‑informed leadership is therefore essential. OECD research on digital government highlights that strengthening the skill mix in public administration is a priority – encompassing both digital competences and what might be termed “data skills for leadership” – the ability of senior decision makers to interpret evidence, ask the right questions of data, recognise its limitations, distinguish correlation from causal evidence, and interpret AI-generated outputs with appropriate caution (van Ooijen, Ubaldi and Welby, 2019[22]).

Furthermore, organisational arrangements in some education systems do not always enable data analysts to contribute effectively. Analysts may be concentrated in small, under-resourced units, distant from policymakers and with limited influence in decision making processes. This can result in the under-utilisation of existing talent. More integrated approaches – such as embedding analysts within policy teams or creating multidisciplinary project groups – can strengthen the relevance of analysis while also improving job satisfaction and retention. This approach is not incompatible with the centralised units or observatories discussed earlier: centralised bodies can provide shared infrastructure and standards, while embedded analysts ensure that findings are applied in day-to-day policy work. Analytical capacity also depends on actors beyond government: an active research community provides an independent source of analysis of public data assets. Where this community is small or poorly connected to available data, valuable datasets risk remaining underutilised.

5.3. Legal and regulatory constraints

5.3.1. Data protection, consent and intellectual property

Legal and regulatory frameworks can shape, and at times constrain, the sharing and use of data for skills policy. Data protection and privacy regulations – such as those introduced under the General Data Protection Regulation (GDPR) – play a critical role in safeguarding personal information, including student records, training participation data and employment histories. However, these frameworks can also make it challenging to link datasets across systems or to use data for evaluation without explicit consent. For example, an education ministry seeking to follow graduates into the labour market by linking education records with tax or social security data may face legal barriers, requiring individual consent or specific exemptions. In many countries, legislation does not explicitly provide for such data matching for policy purposes, leaving potentially valuable datasets disconnected. International evidence further highlights gaps in this area: a recent report found that only 16% of countries explicitly guarantee data privacy in education law, suggesting that in many jurisdictions data protection is either weak – undermining trust – or governed by general laws not tailored to education, creating uncertainty and caution among data holders (UNESCO, 2023[23]).

Consent requirements can also present challenges. When explicit individual consent is required each time data is repurposed for analysis, projects can become impractical. Some countries have sought to address this by enabling the use of anonymised or pseudonymised data for research under strict safeguards, though not all legal regimes allow this without complex approval processes. Intellectual property and licensing issues may add further complications (OECD, 2019[1]). For instance, standardised test data or curriculum materials may be owned by private entities, restricting their sharing or publication. Similarly, private learning platforms may decline to share data with governments, citing intellectual property rights or competitive sensitivities.

5.3.2. Cross-agency and cross-border data sharing

Legal barriers can also arise in cross-agency data sharing. In many administrations, laws or regulations restrict the use of data to the sector in which it was collected – for example, education data may only be used for educational purposes, and employment data only for employment services. Even where such restrictions are not explicit, legal ambiguity can create a “chilling effect”, with public officials reluctant to share data for fear of breaching rules. Without clear legal provisions, for example, a labour ministry may be advised not to share training outcomes with the education ministry. Some countries have sought to overcome these obstacles by enacting specific data-sharing legislation or inter-ministerial agreements that authorise integration for defined purposes, such as monitoring skill outcomes. Where such frameworks are absent, uncertainty and administrative procedures can delay or prevent progress in linking data.

Restrictions on transferring data across jurisdictions present an additional challenge, particularly for international comparisons or collaborative initiatives. Policymakers and researchers may wish to combine data from multiple OECD countries to benchmark performance or identify good practices. However, data localisation requirements and differing privacy regimes often complicate such efforts. Under the GDPR, for example, personal data generally cannot be transferred outside the EU to jurisdictions without equivalent safeguards, which limits sharing with partners in other regions unless the data is anonymised. Even within federations or regional groupings, variations in legislation can impede the flow of skills data.

Public sector data is also subject to freedom of information requirements and confidentiality provisions, which can at times be in tension. Open data initiatives encourage governments to make datasets accessible, while individual-level records on learners and trainees are confidential and require protection. Striking the right balance can be difficult, and in cases of uncertainty, the default response is often to restrict sharing.

5.3.3. Balancing privacy, utility and reform needs

Legal and regulatory frameworks serve a dual purpose: they are essential to protecting citizens’ rights and data security – which is fundamental for public trust – but if outdated or overly restrictive, they can also hinder the effective use of data for policy purposes. The challenge lies in enabling data sharing that is as secure as practicable and lawful, recognising that no system is entirely immune to breaches. Potential approaches include strengthening anonymisation and pseudonymisation techniques, creating legal “safe harbours” for the use of data in statistical and policy evaluation, and updating regulations to reflect the benefits of data integration. OECD work has emphasised the importance of adapting national statistical laws to permit the secure linkage of education, employment and social data for approved research that serves the public interest, under strict privacy safeguards (OECD, 2019[24]). Legal clarity is critical: when agencies understand precisely what is permitted, they can act with confidence rather than defaulting to inaction. Striking the right balance between privacy and utility – ensuring that data are not misused, while avoiding unnecessary barriers to its productive use – often requires legal reform or new regulations. Progress in this area can be slow, which is why legal frameworks continue to represent a major barrier category for effective data use. New Zealand’s Integrated Data Infrastructure (IDI) provides an example of this approach (see Box 3), enabling secure linkage of anonymised data across sectors to support policy analysis while protecting individual privacy.

Box 3. Country example: New Zealand

New Zealand: Integrated Data Infrastructure (IDI)

The Integrated Data Infrastructure (IDI) is a large, anonymised longitudinal microdata resource that links information from multiple sources, including government agencies, official New Zealand surveys, and non-government organisations (NGOs), among others.

By excluding personal identifiers such as names, dates of birth, and identification numbers, the IDI enables cross-sector research while safeguarding individual privacy. This makes it a powerful tool for evidence‑based policy development.

The IDI encompasses eight main categories of data: Health; Education and Training; Benefits and Social Services; Justice; People and Communities; Population; Income and Work; and Housing.

Source: Stats NZ Tatauranga Aotearoa (2022[25]), Integrated Data Infrastructure, www.stats.govt.nz/integrated-data/integrated-data‑infrastructure/.

5.4. Technical and interoperability challenges

5.4.1. Legacy systems

Technical barriers remain a major constraint on the effective use of data in education and skills policy. A key challenge lies in the persistence of legacy IT systems. Many public-sector databases were developed decades ago on mainframe technologies designed for stability but not for interoperability. These systems are costly to maintain, rigid, and difficult to connect with modern applications. By contrast, newer platforms – often based on SQL databases or cloud architectures – allow modular design, easier updates and real-time data sharing.

This mismatch makes interoperability highly complex. For example, linking a database built on a 1990s mainframe with a registry operating on a modern SQL system requires significant technical effort. Governments face considerable challenges when migrating from legacy systems, as these often contain critical data and perform essential functions (OECD Observatory of Public Sector Innovation, n.d.[26]).

Artificial intelligence can help address some of these technical constraints by enabling automated data matching across systems, supporting the harmonisation of different classifications and facilitating the integration of otherwise incompatible datasets. However, these tools remain dependent on minimum levels of data quality and standardisation, and cannot compensate for fundamentally fragmented or poorly structured data environments.

Replacing legacy systems is expensive and carries operational risks. They often contain essential data, and any migration requires careful testing to avoid data loss or service interruptions. At the same time, keeping these systems in place creates long-term problems: older platforms keep data isolated, are difficult to update, and limit the level of interoperability needed for effective policy use. While AI-based tools can support aspects of system integration or documentation of legacy environments, they do not eliminate the need for structural modernisation and investment in interoperable data architectures. Nonetheless, periodic system renewal can also serve as an opportunity to re‑examine administrative processes and the data collected, improving the timeliness and relevance of data collection, though the complexity of such programmes should not be underestimated.

5.4.2. Lack of common standards and definitions

A related challenge is the absence of common data standards and definitions. Education and training systems often rely on different identifiers for the same entities – for example, one database may use a national ID number for individuals, while another uses a student enrolment number, or they may classify fields of study in different ways. Inconsistent definitions create difficulties when datasets are merged, increasing the risk of errors or misinterpretation. For instance, if “completion” of a training programme is defined differently across providers, aggregate statistics on completion rates will lack reliability. A desirable way to address these issues is through the adoption of standardised classifications, such as ISCED for education levels or international occupation codes, which can strengthen data comparability and facilitate linkage across systems. Without such standards, even basic questions, such as how many people received training in a given year, can be difficult to answer consistently, as regions may count hours of instruction in some cases and numbers of participants in others. Some countries have also established sector-specific national data standards; Australia’s AVETMISS standard, for example, defines consistent reporting requirements for vocational education and training data across jurisdictions, supporting comparability and continuous improvements in data quality and timeliness.

Some countries have gone further by creating base registries – centralised databases that provide core information such as individuals, businesses or qualifications. Box 4 illustrates how Estonia addressed these challenges through a fully integrated education information system. The principle is that all government systems reference the same master data, ensuring consistency and avoiding duplication. At the same time, the existence of a single verified personal identifier increases the potential for re‑identification if security is compromised, reinforcing the need for robust and continuously updated privacy safeguards. For example, a national population registry can serve as the single source of verified personal identifiers, which are then used across education, labour and social protection databases. Australia’s Unique Student Identifier (USI) plays a similar role for vocational and tertiary education, enabling training activity to be tracked across providers and over time. A concrete example of the use of standardised identifiers is presented also in, which outlines Chile’s approach to integrating education and labour data.

Box 4. Country examples: Estonfia and Chile

Estonian Education Information System (Eesti hariduse infosüsteem, EHIS)

In 2004, Estonia implemented the EHIS, which has since served as an international reference. This database integrates key components of the education system, including schools, teachers, students, exams, qualifications, and curriculum materials. By connecting directly with schools, it ensures high data quality. Another major advantage is that it enables the longitudinal tracking of students over time. As a result, the data have become more granular, detailed, and regularly updated.

A further efficiency mechanism lies in the EHIS’s connection to other regulated individual databases in areas such as health and employment. This linkage is made possible through the X-Tee – also known as X-Road – a “technological and organisational environment enabling a secure Internet-based data exchange between information systems”.

Chile’s national identifiers (RUT and MRUN)

Chile has developed a robust system of unique identifiers that enables the integration of information across multiple policy domains. The Rol Único Tributario (RUT) is a unique number assigned to every citizen and resident, used primarily for tax and administrative purposes, but also widely adopted across the public and private sectors. The RUT functions as a single identifier that allows data from education, health, social protection and employment systems to be linked reliably to the same individual.

In addition, the education sector developed the Matrícula Rol Único Nacional (MRUN), a unique identifier for students enrolled in the education system. The MRUN is linked to the RUT, ensuring that education records can be connected with other administrative data sources. For example, student progression and completion data can be combined with labour market outcomes captured by social security or tax records. This linkage makes it possible to evaluate the effectiveness of education programmes and to monitor transitions from school to work with greater precision.

The use of RUT and MRUN illustrates how base registries or central reference databases can enhance consistency, reduce duplication, and enable more effective policy evaluation. By relying on a single verified identifier across institutions, Chile has increased its capacity to generate integrated evidence on skills development and outcomes. However, the reliance on a single identifier across systems also heightens long-term privacy risks, underscoring the importance of continuously strengthening safeguards.

Source: OECD (2020[27]), Strengthening the Governance of Skills Systems: Lessons from Six OECD Countries, https://doi.org/10.1787/3a4bb6ea‑en; Information System Authority (2019[28]), Data Exchange Layer X-tee, www.ria.ee/en/state-information-system/data-exchange-platforms/data-exchange-layer-x-tee.

5.4.3. Limited use of APIs and real-time data exchange

Another important technical barrier is the limited availability of application programming interfaces (APIs) and real-time data access mechanisms in education and training systems. APIs are tools that allow different databases and applications to communicate with one another in a secure and automated way. In practice, this means that data held in one system – for example, student performance records – could be directly linked with information from another, such as local labour market indicators, without the need for repeated manual extraction and cleaning. When APIs are missing or underdeveloped, data integration becomes slow, labour-intensive and prone to errors, often requiring staff to download, reformat and merge files from multiple sources. This reduces the timeliness and reliability of the information available to policymakers.

More regular and automated flows of information would be particularly valuable for monitoring key indicators such as employability outcomes of VET graduates, early school leaving and dropout rates, enrolment in lifelong learning programmes, or the alignment of training provision with local labour market demand. Without such mechanisms, valuable data remains underused, and analytical capacity is diverted to data preparation rather than policy analysis. APIs themselves also require governance: mechanisms to validate use cases, define interface contracts and ensure that protected data are shared only under appropriate conditions. More broadly, interoperability should encompass data traceability and lifecycle management, so that when data are transformed or combined to produce indicators, their origin and the transformations applied remain auditable.

5.4.4. Cybersecurity and data resilience

Cybersecurity represents a critical challenge in the development of data-driven skills systems. As databases become increasingly interconnected, their exposure to cyber risks also grows (OECD, 2023[14]; OECD, 2023[29]). A breach of sensitive information – such as student or trainee records – could undermine public trust and significantly delay progress in data integration efforts. Concerns about such risks often make policymakers and IT departments cautious about linking systems, fearing that broader access may increase vulnerability. Insufficient protection against cyberattacks remains a pressing issue, with potential legal, operational and reputational consequences.

Data resilience poses a distinct set of challenges. Many older systems lack robust backup and disaster recovery mechanisms, creating a risk of prolonged downtime or even data loss in the event of system failure, whether caused by technical faults, natural disasters or human error. Building resilience – ensuring that data systems can withstand disruptions and recover rapidly – is essential for the continuity of evidence‑based policymaking. If a key education or labour database were to become unavailable during a critical period, such as budget planning, it could directly compromise decision making. Addressing resilience requires sustained investment in infrastructure redundancy, regular testing of recovery procedures and clear protocols for maintaining analytical capacity during disruptions.