Executive summary
Drawing on the Longitudinal Educational Outcomes (LEO) dataset for England, this analysis investigates how educational choices and prior attainment influence earnings for disadvantaged students, highlighting implications for policy and careers guidance.
Early career earnings are strongly associated with post-16 qualifications
- Earnings at age 25 are strongly associated with the qualifications taken post-16. Those with level 3 qualifications earn substantially more on average than those with lower-level or no qualifications.
Disadvantaged students go on to earn less on average than their non-disadvantaged peers
- Regardless of the post-16 qualifications taken, disadvantaged students aged 25 consistently earn less on average than their non-disadvantaged peers. For example, male disadvantaged students who took level 3 qualifications earned, on average, £3,250 less than their non-disadvantaged peers (£25,020 v £21,770).
- GCSE attainment at 16 is a key driver of this earnings gap, especially for women. The raw earnings gap between disadvantaged and non-disadvantaged students is around £4,100 but falls by £2,000 for women and £1,500 for men after controlling for prior attainment.
The lower prior attainment and qualification pathways of disadvantaged students explain much of the earnings gap
- Much of the impact that qualification choices seem to have on the earnings gap between disadvantaged and non-disadvantaged students is actually due to differences in GCSE results. Students’ GCSE results strongly influence the post-16 routes they take, and those routes in turn affect earnings.
- However, qualification choice does appear to make some contribution to the disadvantage gap, over and above the impact of GCSE attainment, accounting for at least 8 per cent of the earnings gap for women and at least 3 per cent for men.
- A large proportion of the earnings gap remains unexplained (i.e., not attributable to GCSE attainment or qualification level). We estimate this proportion to be around 38 per cent for women, but considerably larger at around 62 per cent for men.
Methods
Using linked education and employment data from the Longitudinal Educational Outcomes (LEO) dataset, we analyse the employment status and yearly earnings of students as they enter the labour market following their study.
We review the effect of qualification choice and prior attainment on pre-tax earnings at age 25 for male and female disadvantaged and non-disadvantaged students, covering the 2017-18 to 2020-21 tax years. This represents the latest three-year period for which earnings data is available in the LEO dataset. Earnings are averaged across this period, with all figures expressed in 2020-21 prices using HM Treasury GDP deflators.
Only students whose employment can be linked to education data are included, excluding, for example, students who left the country or never entered employment. We have excluded the top and bottom 5 per cent of earners in each tax year to remove outliers, removing individuals who receive little-to-no earnings and those who earn vastly higher figures than the greater population at age 25. Finally, qualifications are grouped into levels 1-3 following DfE definition of qualification levels.[1]
The early career earnings of disadvantaged students by qualification level
Before analysing the contributing factors to any earnings gaps, we first review the scale of earnings gaps between disadvantaged and non-disadvantaged students themselves, and how these differ by qualification level.
Figures 1.1 and 1.2 show our average earnings measure for female and male students respectively, split by disadvantage and qualification level.
We immediately observe raw differences in earnings across all characteristics. Those who achieved level 3 qualifications[2] earn the most on average, followed by level 2 qualifications and level 1 and no qualification. Disadvantaged individuals and women also earn less at each qualification level.
Figure 1.1: Average raw earnings of female students at age 25 by disadvantage status and highest qualification entered by age 19 (2017-18 – 2020-21, averaged)
Figure 1.2: Average raw earnings of male students at age 25 by disadvantage status and highest qualification entered by age 19 (2017-18 – 2020-21, averaged)
What is the association between post-16 qualifications and early career earnings?
As we have seen, there are clear raw differences in average earnings between disadvantaged and non-disadvantaged students. While there are likely many factors that influence these differences, what we aim to investigate in this analysis is the effect that post-16 qualification choice, and the corresponding socio-economic differences in qualification choice observed above, has on a student’s later earnings.
To do this, we have used a series of linear regression models to estimate the degree to which the observed disadvantage gap in earnings remains after accounting for relevant factors. These models are shown separately for male and female students, and are constructed sequentially as follows:
- Model A: controlling only for disadvantage at key stage 4.
- Model B: controlling for disadvantage at key stage 4 and GCSE attainment.
- Model C: controlling for disadvantage at key stage 4, GCSE attainment, and highest qualification entered by age 19 (up to Level 3)
- Model D: controlling for disadvantage at key stage 4, GCSE attainment, highest qualification entered by age 19 (up to Level 3), employment region, and employment industry.
In this sequence, Model A represents the ‘raw’ earnings gap between disadvantaged and non-disadvantaged students at age 25. Model B shows the gap after accounting for differences in GCSE attainment, while Model C shows the gap after additionally accounting for differences in qualification choice. Model D adds controls for region and industry of employment. The models are not intended to provide a full account of all possible influences on earnings, but rather to give an indication of the extent to which GCSE attainment and post-16 qualification choice are associated with the observed disadvantage gap in earnings.
Figure 1.3 shows the size of the disadvantage coefficient, or earnings gap in £, for our series of models for female and male students. For both female and male students, we can see that accounting for GCSE attainment significantly reduces the earnings gap. For female students, the earnings gap is halved (~£2,000) after accounting for prior attainment, while for male students this effect is smaller (~£1,500). This indicates the lower GCSE attainment of disadvantaged students is a significant driver of their lower earnings, especially for female students.
Adding post-16 qualification entries into the model only has a modest impact on the gap. This reflects the fact that prior attainment and qualification entries are closely linked – post-16 qualification choices are often heavily constrained and influenced by student’s Key Stage 4 GCSE results, and so much of the effect of qualification choice on earnings is already captured in the GCSE attainment term. In other words, the smaller change between Models B and C does not mean that qualifications have little impact on earnings, but rather that much of their impact operates indirectly via GCSE attainment. The additional reduction for females suggests that qualification choice may still play a somewhat larger role in the earnings gap for women than for men.
Following the addition of employment region and industry, we observe a slight decrease in the size of the earnings gap for male students. This suggests that disadvantaged male students are more likely to work in lower-paying regions and industries than their peers with the same GCSE attainment and qualifications. For female students on the other hand, the gap shows no statistically significant change. This suggests that, unlike for male students, the earnings gap for female students cannot be further explained by disadvantaged students being more likely to work in lower-paying regions and industries than their non-disadvantaged peers with the same GCSE attainment and qualifications.
Figure 1.3: Size of disadvantage earnings gap at age 25 across models, (2017-18 – 2020-21, averaged)
Breaking down the earnings gap
To further understand contributing factors to the earnings gap, we have used an Oaxaca-Blinder decomposition.
Interpreting our decompositions
The Oaxaca-Blinder decomposition is a statistical method that ‘decomposes’ (or breaks down) differences in average outcomes across two groups into distinct kinds of contributions:
- Explained contribution – this is where differences in the characteristics of disadvantaged and non-disadvantaged students contribute towards the gap. For example, as disadvantaged students tend to have relatively lower levels of prior attainment, differences in prior attainment are likely to contribute to the explained part of the gap (as seen above). We have split this part of the gap into the two factors modelled in this analysis: prior attainment and qualification level.
- Unexplained contribution – this is where the effects of a factor appear to differ between disadvantaged and non-disadvantaged students. Practically, this means that although some portion of the gap can be explained due to differences in prior attainment or qualification level, the effect these factors have on earnings is greater for disadvantaged students than it is for non-disadvantaged students. This unexplained portion also includes the part of the gap than cannot be explained by the factors we have modelled, such as parental education levels.
- Interactions – this, often smaller, part of the gap is the contribution made by disadvantaged students and non-disadvantaged students both having different characteristics, and those characteristics being rewarded differently. In our models, this effect is much smaller than other observed effects.
Figure 5.4 presents two separate Oaxaca–Blinder decompositions for female and male students: one including both GCSE attainment and highest qualification, and one including only GCSE attainment.
Figure 5.4: Decomposition of the disadvantage earnings gap at age 25, 2017-18 – 2020-21, expressed in 2020-21 prices
Focussing on female students first, when both GCSE attainment and highest qualification are included together, they explain 56.2 per cent of the earning gap.[3] When GCSE attainment only is included it only explains 48.5 per cent of the gap. This indicates that qualification choices do contribute to the disadvantage earning gap. Specifically, disadvantaged students who go on to choose lower-level qualifications than their non-disadvantaged peers, even when they have the same GCSE results, subsequently earn less. This is consistent with previous research which suggests that disadvantaged students, particularly women, are more likely to choose to study courses in post-16 education associated with lower earnings in later life.[4]
For male students, there is a similar pattern, although the effects are smaller in magnitude. When both GCSE attainment and highest qualification are included together, they explain 39.8 per cent of the earning gap. When GCSE attainment only is included, this falls to 36.9 per cent of the gap. Again, there is some indication that qualification choices make a contribution to the earnings gap, but this effect is considerably smaller for males than it is for females.
For both genders, the interaction term – representing the combined effect of disadvantaged students having different characteristics and those characteristics being rewarded differently – is relatively small. It is slightly larger and consistently positive for females, suggesting a mild compounding effect of disadvantage in which different prior attainment and qualification patterns are also less well-rewarded. For males, the interaction term is smaller and sometimes negative, indicating little or no compounding effect.
Overall, the decompositions reinforce two points. First, attainment at GCSE is a significant driver of the disadvantage earnings gap for both genders, with a large part of its effect mediated through subsequent qualification choices. Second, qualification level appears to make a larger direct contribution to the gap for women than for men, consistent with earlier research linking early-career gender pay differences partly to subject and qualification choices.[5]
Conclusion and recommendations
Considering labour market outcomes, English data shows that GCSE attainment explains a substantial share of the disadvantage-related earnings gap at age 25, with qualification choice adding only a small additional contribution once prior attainment is accounted for. This pattern reflects the close link between attainment at age 16 and access to higher-return qualifications, underscoring the importance of narrowing attainment gaps early in education.
However, the decomposition results also indicate that post-16 qualification patterns have a more pronounced association with earnings for female students. For disadvantaged women, differences in qualification level appear to contribute more to later earnings gaps than for men, suggesting that route choice and progression pathways may play a larger role in shaping their early career outcomes. This highlights the importance of ensuring that careers information, advice, and guidance (CIAG) effectively supports disadvantaged young women to access higher-level and higher-return qualifications, and that barriers to progression—such as perceptions of subject suitability or access to provision—are addressed. Strengthening CIAG and widening opportunities for progression into both academic and technical level 3 routes could help ensure that equally able disadvantaged students, particularly women, can realise their full potential in education and the labour market.
