Qamar Zaman
Partner, Stripe Partners
Our Revealing Social Capital project funded by the Nuffield Foundation, explores how social connections influence people’s life outcomes, with a focus on social mobility. As part of our ongoing research, we’re publishing, in collaboration with Stripe Partners, a series of regional social mobility statistics for England. Later this year, we plan to publish more detailed social mobility estimates, covering smaller geographical areas and using different approaches.
We use Longitudinal Education Outcomes (LEO) data which links individuals’ school records with their adult earnings data to construct a simple measure of social mobility (based on van der Erve et al. (2024)). As a proxy for parents’ income during a person’s childhood, we use whether someone was eligible for Free School Meals (FSM) at age 16. FSM eligibility is a widely used indicator of economic disadvantage, as it typically signifies parents in the lowest income groups. Data on FSM eligibility at age 16 for the 1986-92 birth cohorts allows us to identify the ~14% of individuals from the lowest-income families. We then observe their earnings rank at age 28, using a higher rank as an indicator of greater upward social mobility relative to their disadvantaged starting point.
By age 28, significant earnings gaps exist between those who were eligible for free school meals (FSM) at 16 and those who were not. Men who were FSM-eligible earn a median annual income of £13,753, which is 37% less than their non-FSM peers (£21,771). For women, this gap is even more pronounced: FSM women earn £6,644, merely two-fifths of what non-FSM women earn (£16,187). However, we must interpret the data for women cautiously. The lack of information on working hours limits our analysis. Research shows that family background influences the timing of having children. These different fertility patterns often result in women from disadvantaged backgrounds engaging in more part-time work at age 28 compared to their peers. This employment pattern makes it challenging to draw accurate conclusions about earnings disparities among women based solely on our data.
We’ve examined these social mobility statistics across 326 local authorities in England to understand regional differences. Our analysis highlights areas with the highest and lowest social mobility rates. You can explore these variations on the interactive map below or download the full dataset here.
The place where disadvantaged children grow up significantly influences their future outcomes. Median earnings for FSM men at age 28 are particularly low in major northern cities and towns such as Newcastle or Leeds. In contrast, several areas in the South East show high social mobility for boys, including parts of Greater London (like Tower Hamlets) and nearby areas (such as Hertfordshire). However, it’s not a simple north-south divide. Pockets of low social mobility exist across southern England. Within London, earnings at age 28 for children who received free school meals vary considerably between boroughs.
Using the latest iteration of LEO, we were able to extend van der Erve et al.’s (2024) work by covering seven birth cohorts (1986-1992) instead of three (1986-1988), analysing 2 million men and 1.9 million women. Our findings using additional birth cohorts closely match van der Erve et al.’s results, with identical average ranks for FSM men, and near-identical average ranks for FSM women. The median earnings for FSM and non-FSM eligible individuals differ somewhat, but the overall picture remains consistent.
Men | Women | |||
Birth cohorts | 1986-88 | 1986-92 | 1986-88 | 1986-92 |
Average rank of FSM children | 42.2 | 42.2 | 32.0 | 32.4 |
Median earnings of FSM children | £12,513 | £13,753 | £6,068 | £6,644 |
Median earnings of non-FSM children | £20,573 | £21,771 | £15,095 | £16,187 |
At the local authority level, our estimates of social mobility show little change too. When comparing calculations using three-year versus seven-year cohorts, the average fall or increase in FSM children’s income positions relative to their non-FSM peers is only 1.3 rank points across local authorities. This suggests that social mobility for more recent birth cohorts in England remained relatively unchanged, even after taking regional differences into account.
The second way the new LEO data allows us to extend previous analyses is by exploring social mobility for the oldest three birth cohorts using earnings at age 32 instead of just 28. This approach helps address concerns about early career earnings potentially misrepresenting lifetime earnings.
As expected, our findings for men show that median incomes are higher at 32 than at 28 for both FSM and non-FSM-eligible individuals. However, the average income rank of previously FSM-eligible men at 32 is only slightly higher (1.6 percentiles) than at 28. For FSM women we observe an even smaller increase in their average income rank between ages 28 and 32. The figures for non-FSM women highlight the previously referenced limitations of these data. We see a decrease in their median earnings, likely due to reduced working hours, possibly from family-related responsibilities. The data for FSM women does not show this earnings decrease, potentially because their work patterns were already characterised by fewer hours at age 28.
Men | Women | |||
Age | 28 | 32 | 28 | 32 |
Average rank of FSM children | 42.2 | 43.8 | 32.0 | 32.6 |
Median earnings of FSM children | £12,513 | £15,197 | £6,068 | £6,682 |
Median earnings of non-FSM children | £20,573 | £23,575 | £15,095 | £14,748 |
At the local authority level, our estimates of social mobility show little change too. When comparing income ranks at ages 28 and 32 for FSM men, the average difference is only 1.9 rank points across local authorities. This small change suggests that relative income positions established by age 28 remain largely stable into the early 30s, even when accounting for regional variations.
Research in the US has shown that social mobility can vary significantly even between small, neighbouring areas. Building on this insight, our next steps are twofold: Firstly, we plan to calculate and release social mobility estimates for areas smaller than local authorities. This will provide a more nuanced and granular picture of geographical variations in mobility. Secondly, we’ll expand our approach to measuring childhood socioeconomic status. Instead of relying solely on FSM eligibility, we’ll test other methods to estimate parental income during a child’s upbringing. This will allow us to consider a broader spectrum of economic backgrounds when estimating social mobility. We hope these further analyses will offer a more comprehensive understanding of social mobility patterns in England.
This work contains statistical data from ONS which is Crown Copyright. The use of the ONS statistical data in this work does not imply the endorsement of the ONS in relation to the interpretation or analysis of the statistical data. This work uses research datasets which may not exactly reproduce National Statistics aggregates. All analyses have been conducted in the ONS Secure Research Service.
Partner, Stripe Partners
Junior Data Scientist, Stripe Partners
Advisor
Principal Advisor
Design and development by Soapbox.