Better skills data for smarter financing of education and training

Public investment in skills – used here broadly to encompass the knowledge, competencies and qualifications acquired through education and training, from early childhood to adult learning – is a cornerstone of economic and social development. Higher levels of education and training are consistently associated with better employment outcomes, higher earnings, and stronger social cohesion, though returns vary across pathways, population groups and context. Governments thus see skills spending not merely as a recurrent cost but as an investment in future prosperity.

Approaches to financing skills development vary considerably across countries and stages of the life course. Each system combines a particular mix of financial instruments – ranging from supply-side funding of institutions to demand-side mechanisms such as subsidies or vouchers – shaped by national priorities and institutional contexts. The effectiveness of these arrangements depends on their alignment with policy goals and on the quality of the underlying evidence base. Efficient resource allocation has become increasingly important, not only in terms of how much is spent but in ensuring that funds are directed to interventions with proven impact.

At the same time, governments operate under growing fiscal constraints. Ageing populations, rising health and social care costs, climate adaptation, and new security pressures continue to exert pressure on public budgets. In this constrained environment, skills investment must compete with other critical priorities. Redirecting resources from inefficient or low-impact programmes proves politically challenging, as entrenched interests and institutional inertia resist change. However, tighter fiscal conditions make evidence‑based decision making indispensable. It enables governments to justify, sustain and, where appropriate, expand investment in skills development by demonstrating clear value for money and measurable outcomes.

Across the OECD, the lack of timely, reliable and integrated data is a major barrier to strategic decision making in skills investment. Governments often lack comprehensive information on the effectiveness of education and training programmes, the needs of different population groups, and the long-term labour market outcomes of participants. While relevant data sources are often available – from administrative records and employer surveys to tax and employment registers – their coverage, quality and accessibility vary significantly across countries. In many cases, data remain fragmented, incomplete, or difficult to access and use for policy purposes.

OECD work on the data-driven public sector shows that these challenges are not unique to skills policy. Governments across policy domains face similar difficulties, including uneven data quality, limited internal data sharing, and barriers to interoperability that prevent data from being used as a strategic asset for policymaking and service improvement (OECD, 2019[1]). However, these challenges are particularly acute in skills policy. Skills systems span multiple ministries, levels of government and stakeholders, from education and training providers to employers and social partners. This institutional complexity makes it more difficult to assemble the information needed to assess outcomes across the life course and to link spending with results. OECD analysis also highlights the breadth and diversity of stakeholder engagement arrangements in skills systems, further underscoring the co‑ordination challenges involved (Reznikova, Labanino and McKee Mathews, 2024[2]). As a result, gaps in data integration and use have more direct consequences for the allocation and effectiveness of public investment in skills.

Without quality data, governments cannot determine whether skills investments are generating the intended outcomes or how to target resources for maximum effect. This leaves them vulnerable to inefficiencies and missed opportunities. Funds may flow to programmes with weak results – even where this reflects deliberate policy choices to support disadvantaged groups – while high-impact initiatives remain underfunded. Insufficient data have thus become a critical bottleneck, preventing the alignment of skills policies with economic and social needs.

It is important to recognise that governments are not the sole actors in developing and deploying skills. Employers, private providers and individuals invest significantly in training, often responding more rapidly to evolving labour market demands than public systems can. Better public data should complement – not substitute for – these private contributions.

Better data would transform policymaking. Consider the questions governments currently struggle to answer. Which education and training pathways deliver lasting employment gains? Which sectors and regions generate the highest returns from upskilling? How do different forms of financing affect participation and outcomes across the life course and across different population groups, particularly those at risk of exclusion? With reliable evidence, governments can make more strategic choices. They can invest where impact is greatest, adjust programmes to meet evolving labour market demands, and ensure resources are used efficiently, effectively and equitably – goals that are complementary rather than competing in the context of public skills financing. Advances in digital tools, including artificial intelligence, further expand these possibilities by enabling more granular analysis, improved forecasting of skills needs and more responsive allocation of funding. However, their effectiveness depends critically on the availability of high-quality, integrated data.

While closing data gaps has significant potential to strengthen policy decision making, doing so requires overcoming a set of structural, institutional and technical barriers. These challenges are diverse and operate at different levels of the system, shaping how data are collected, governed, shared and used in practice. To support a clearer and more action-oriented understanding, this paper proposes a framework that groups the barriers into four categories: institutional, governance and financing barriers; human-capital and analytical-capacity gaps; legal and regulatory constraints; and technical and interoperability challenges. Figure 1 illustrates these four categories, summarising what each looks like in practice and why it matters for decision making and the financing of skills systems. Each category involves distinct implications for policy design and requires different levers for reform. By structuring the analysis in this way, the section aims to provide a practical basis for identifying priorities and sequencing actions that enable progress toward more integrated and effective skills data systems. Throughout this paper, “skills” refers broadly to the knowledge, competencies and qualifications acquired through education and training, consistent with its use in international policy discourse.

This policy paper examines how stronger data on skills outcomes and investments can enable smarter public spending decisions, and what obstacles currently impede this goal. While many of the data issues discussed are relevant to skills policy more broadly, the paper’s focus is on data insofar as they inform financing and resource‑allocation decisions: assessing programme effectiveness, targeting investment and linking spending to outcomes. Section 2 identifies current data gaps in education and skills policy and the consequences of those gaps for decision making. Section 3 provides an overview of the main data sources available for education, employment and skills policy, outlining what each source can and cannot support in terms of evidence and decision making. Section 4 discusses what is at stake – the potential gains in efficiency and socio‑economic outcomes if these gaps are addressed. Section 5 analyses the main barriers to progress, including institutional, governance and financing constraints, legal and regulatory challenges, capacity limitations and technical interoperability issues. Finally, Section 6 presents the conclusions.