Understanding labour markets through administrative data: A workshop of collaborating minds

This blog is by ADR UK Research Fellow Guglielmo Ventura, who reflects on a recent workshop hosted at the LSE Centre for Economic Performance. The event brought together leading academic and policy voices to explore how linked administrative data can deepen our understanding of young people’s labour market transitions and inform better policy.

In November 2025, the LSE-based Centre for Economic Performance (CEP) hosted a workshop as part of my ADR UK fellowship, “Unpacking returns to vocational education: The role of skill, labour market sorting and mobility”. The event featured leading academics presenting cutting-edge research using linked administrative data, followed by a focused panel discussion on how we could better use available data to fill knowledge gaps around the labour market challenges faced by young people in the UK.

The talks were insightful and wide-ranging, both in topic and geographic scope. They included, among the topics:

• research on the work experiences of less-educated workers and how social skills can help them thrive in the workplace
• the role of labour market conditions at entry in driving qualification mismatch in Germany
• evidence from Norway on how the shift towards non-routine job tasks affects students’ education choices and social mobility
• analysis from Sweden on the impact of degree choice on fertility patterns.

Richard Blundell (UCL and Institute for Fiscal Studies), Alex Collinson (Trades Union Congress), Becci Newton (Institute of Employment Studies) and Anna Round (Youth Futures Foundation) joined the panel discussion, chaired by Sandra McNally (CEP). The discussion reminded us that over one million young people in the UK are currently not in employment, education or training (NEET). Falling vacancies and rising unemployment risk pitting young people who are NEET against a growing pool of experienced jobseekers, exacerbating the problem of long-term worklessness among young people.

Some groups of young people are more at risk than others. School leavers without good grades in their GCSEs at 16 face a higher likelihood of being out of work or education, and a disadvantaged socio-economic background or special educational needs can further exacerbate this risk. Young people who interact with the care or justice system are particularly at risk too. More generally, those who do not follow conventional and clearly signposted education pathways are often short-changed by the system, with negative consequences for social mobility.

Higher rates of worklessness and a lack of education or training are not the only challenges. Low-paid employment and anaemic pay growth disproportionately affect young people, particularly in certain parts of the country. While these problems are widely recognised, understanding their underlying causes, and identifying effective policy solutions, is a more elusive task.

What we know and don´t know

After many years lagging behind administrative data powerhouses elsewhere in Europe, the UK has seen tremendous improvements in research access to linked administrative data over the past decade. Thanks to the linkage between education records and tax and benefits data, the Longitudinal Education Outcomes (LEO) dataset has transformed researchers’ ability to follow young people as they move through education and into the labour market.

With full coverage of the school-leaver population, LEO allows researchers to credibly study transitions for disadvantaged groups that survey-based studies often struggle to capture due to small sample sizes. By comparison, the Annual Survey of Hours and Earnings (ASHE)—for many years the closest thing the UK had to a linked employer–employee dataset and a key source of information on workers´ pay—is based on a 1% subsample of the population. This makes it too limited for studying specific at-risk groups in detail (more details on ASHE can be found on the ASHE linked to PAYE and Self-Assessment data page).  

Despite this leap forward, the panel discussion laid bare how much we still do not know about youth labour market transitions. It also made clear that further improvements in data collection, linkage and access are essential if we are to close these knowledge gaps. Foremost, if we are to help every young person succeed in the school-to-work transition, we need to develop a much deeper understanding of the job matching process. What is it that makes a job “good”? Who has access to “good jobs”, and through what mechanisms?

Datasets such as LEO allow us to identify whether young people are working for “good” or “bad” firms. My own ADR UK-funded research explores how this shapes differences in labour market trajectories by education. Ideally, however, we would also like to understand whether firm quality depends on the occupations offered, or on the types of co-workers from whom young people can learn. We also know very little about the skills demanded by local firms — demands that are themselves shaped by business decisions, especially when it comes to technology adoption.

Future improvements

There is a clear need for richer information on both employers and workers. LEO is illuminating when it comes to the qualifications obtained by young people, but has little to say on the actual skills young people acquire. As a result, we do not know whether they possess the right skills to match local employers´ needs. Being able to say something more on this is not wishful thinking. Many countries´ linked administrative data includes information on occupations which can be linked with information on task requirements, which was an approach highlighted in several workshop presentations.

Thanks to ADR UK, research access to administrative data is continually expanding, and future linkages promise to add greater depth to analyses of youth labour market transitions. For example, linking LEO with the Millennium Cohort Study would introduce much richer qualitative measures. In this respect, improving the ability to combine administrative data with other data sources remains crucial to weave in the voices and lived experiences of the communities the data is about, thus fully realising its public benefit.

Importantly, panel members’ and audience’s contributions alike emphasised that while ADR UK’s flagship datasets are already enabling researchers to answer critical questions, they are also generating demand for access to additional information. This virtuous cycle, where new evidence leads to deeper questions, ultimately supports the expansion of knowledge and strengthens research in the public good.