Returns to academic and vocational qualifications at 16: new evidence from linked administrative data

Working Paper
Using the Longitudinal Education Outcomes (LEO) dataset linked to the National Pupil Database, this paper estimates the earnings returns to GCSE grades and Level 2 vocational qualifications and finds substantial variation by subject, gender, and prior attainment.
Author

A. Smith, B. Jones, C. Williams

Published

May 1, 2026

EERC Working Paper · May 2026

Abstract

Using the Longitudinal Education Outcomes (LEO) dataset linked to the National Pupil Database (NPD), this paper estimates the earnings returns to academic and vocational qualifications at age 16. We exploit variation in the timing of qualification reforms to identify causal effects, and find that returns to academic qualifications are, on average, positive and significant for both men and women. Returns to Level 2 vocational qualifications are more heterogeneous, with meaningful variation by subject area, awarding organisation, and learner background.

Our results suggest that headline comparisons of vocational and academic pathways conceal considerable within-category variation. We discuss implications for how qualifications policy and careers guidance should be calibrated to reflect differences in labour market value across routes.

Key findings

  • The average earnings premium to five or more good GCSEs (grades 4+) is approximately 8–12% at age 25, conditional on subsequent qualifications.
  • Returns to Level 2 vocational qualifications vary from close to zero (in some service sector subjects) to around 6–9% (in construction and engineering).
  • Gender gaps in returns are pronounced in vocational subjects, reflecting patterns of occupational segregation.
  • Pupils from lower socioeconomic backgrounds are less likely to take qualifications with higher labour market returns, conditional on ability.

Data and methods

The analysis uses matched NPD–LEO data covering school leavers between 2005 and 2015, with earnings outcomes observed at ages 23–30. We use an instrumental variables strategy based on distance to providers offering specific qualification types.

Download working paper (PDF)