How do we address the gap in attainment between the most advantaged and disadvantaged students in the UK? Pioneering research, led by the University of Bristol, reveals the reforms most likely to equalise our education system.
Lead author Simon Burgess, Professor of Economics, explains how the team’s findings could lead to much-needed changes in how school places are allocated.
There is much to applaud about the school system in England, but also deep problems. Chief among these is the wide and persistent gap in educational attainment between disadvantaged children and pupils from more affluent families.
For example, in 2019, around 30% of pupils eligible for Free School Meals (FSM) achieved the benchmark performance in GCSEs, compared to double that among more affluent pupils. This gap has barely changed for at least 20 years.
Part of this gap arises from differences in the effectiveness of the schools these children attend. Richer pupils are much more likely to be assigned to effective secondary schools.
In fact, richer pupils are over 40% more likely to attend a highly effective secondary school (in the top 25% of value-added, in England called Progress 8). Not only might this be considered unfair for the current generation, it can also perpetuate income inequality through the generations.
The geography problem
Differences in the effectiveness of schools attended might simply be the result of families’ preferences for schools. Our research, however, shows that admissions arrangements play an important role in explaining the observed unequal attendance at effective secondary schools.
Specifically, most English secondary schools explicitly prioritise pupils according to where they live – either through defined catchment areas or by ranking applicants by straight-line distance between home and school.
This is not neutral: desirable schools generate substantial house price premiums in their catchment areas, effectively pricing out lower-income families. School choice through residential location appears not to be an option for poorer families. We show that richer pupils disproportionately move into the catchment areas of popular schools during their primary school years.
Our research – analysing over 550,000 pupils and 3,248 schools across 152 local authorities – provides the first comprehensive picture of this dynamic. We assembled detailed data on individual school admissions arrangements (a significant undertaking given the absence of any central record) and linked these with families’ expressed school preferences, locations, and socioeconomic characteristics.
We first looked at the schools within a commutable range of pupils’ homes and calculated the average effectiveness of those schools for each pupil. Then we factored in schools’ geographic admissions criteria and looked at the average effectiveness of the schools that each pupil could access in practice.
The pattern by income disadvantage is stark: the gap between schools accessible by commute and in practice is much higher in poorer neighbourhoods. In other words, the effectiveness penalty of imposing the geographic criteria worked strongly against FSM-eligible pupils.
We also documented inequality in educational access across English Local Authorities (LAs). LAs with greater prevalence of geographic admissions criteria have significantly and substantially higher levels of inequality in access to effective schools.
A targeted intervention
Our aim is to weaken the power of geography in school admissions, by giving priority to other groups for some school places. Weaken, but not totally remove, since geographic criteria have advantages too: they create a sense of community and make active travel to school (walking/cycling etc) feasible.
We modelled three potential reforms to school admissions criteria. Briefly, these are: priority for FSM-eligible pupils up to a quota; a quota of open places, allocated using a ballot; and banding, a test-based process intending to give all schools similar ability profiles.
These are all designed to make marginal changes to the system, and indeed all the simulations found that approximately 90% of pupils would attend the same school as under current arrangements.
Of the three reforms we analysed, the admissions priority for FSM-eligible pupils, up to a quota, demonstrates clear superiority in both effectiveness and feasibility.
The idea is straightforward. Each school would reserve 15% of places for FSM-eligible pupils, with remaining places allocated according to its existing admissions arrangements. The idea is that the reform tempers geography with a measure of social equity.
Our simulation results, combining families’ estimated school preferences with school admissions criteria, indicate a substantial effect. Under this reform, virtually all FSM pupils would be assigned to one of their top two preferences – typically representing a more effective school than their school assignment without reform (baseline).
The average effectiveness of schools to which FSM-eligible pupils are assigned is higher by 15.9% over baseline, and the gap in average school effectiveness between disadvantaged and advantaged pupils falls by 17.0%. And yet, 94% of pupils are assigned to the same schools as baseline.
This reform is therefore very effective and very targeted.
Anticipated objections
‘Why not just make all schools great?’ Of course, that would be the best thing to do. But decades of effort have shown that systemic, universal school improvement is extraordinarily difficult to achieve, and we have not yet managed it. In the meantime, while we work on that longer-term goal, we can try to make access to effective schools fairer.
‘Home background matters far more than schools.’ It is indisputable that within-school variation in outcomes exceeds between-school variation – home circumstances matter hugely. But that point answers the wrong question, it’s too abstract. We also know that trying to equalise home circumstances, specifically household income, has proven politically intractable. The immediate practical question is: what policies can make the most difference to pupil outcomes? The policy we propose is administratively straightforward and very effective.
‘Most FSM students get their first-choice school anyway, so very little will change.’ This misunderstands constraints on families’ submitted school choices. When families know they have no realistic chance of admission to better schools, they don’t waste their limited choices applying there; typically, people apply to where they can get in. Our reform considerably expands the feasible set of schools for FSM-eligible pupils, meaning disadvantaged families can aim higher. When accounting for true preferences, we find that our reform increases the fraction of FSM pupils getting their first choice school from 91% to 98.6%.
‘What about school community?’ Geographic proximity undoubtedly fosters community ties. Children can both learn and play together, homework clubs and so on are much easier. But this cohesion comes at the cost of excluding other families from the school. There’s a trade-off between community and equity. Current policy sits at one extreme, heavily favouring community while ignoring equity. Our proposal shifts the balance modestly towards equity while preserving most of the community benefits: 85% of places would still be allocated by the existing admissions system.
‘Won’t this bring school run traffic chaos?’ Analysis of baseline travel patterns demonstrates that proximity to home is already imperfectly correlated with school attendance – many pupils travel past their nearest school. Our modelling indicates minimal increases in average journey distance under the reform, with the change in the median distance of just 24 metres.
‘Won’t house prices fall?’ Some house price falls in premium catchment areas are plausible, given that current premiums partially reflect enhanced school access. However, because geographic criteria would continue to allocate most places, households in immediate proximity to desirable schools would retain very high admission probabilities. Anticipated price effects would therefore be modest and localised rather than systemic.
Making it happen
We believe in evidence-based policy and would like to see our preferred policy implemented at the national level, through the School Admissions Code. However, before making this step we would suggest a randomised controlled trial across several local authorities.
Ideally, this would be a one-year trial to allow rapid evaluation of changes in assignment rather than delaying change for years.
However, the case for intervention is already empirically robust. Geographic admissions criteria generate and sustain educational access inequality through well-documented mechanisms. Priority for FSM-eligible pupils up to a quota would materially narrow the gap in effectiveness in school assignments, expanding opportunities for pupils from disadvantaged backgrounds, currently excluded from high-performing schools.
The reform would make a substantial difference to attainment and, beyond school, to subsequent life chances.
It has become a political commonplace that your life chances should not depend on your postcode. And yet, that is the very system we have in our schools, kept in place by successive governments for decades. Adopting this targeted and effective reform can start to change that.
Simon Burgess, University of Bristol and IZA.
Estelle Cantillon, Université Libre de Bruxelles, FNRS, and CEPR.
Mariagrazia Cavallo, Luxembourg Institute of Socio-Economic Research (LISER).
Ellen Greaves, University of Exeter and IZA.
November 2025
This blog post is based on our research report – all details and definitions can be found there.
We are very grateful to the Nuffield Foundation for funding the project and to the UK Department of Education for providing much of the data via the Office of National Statistics’ (ONS) SRS facility.
This does not imply the endorsement of the ONS or other data owners. This does not imply ONS’ acceptance of the validity of any methods used or the output itself, and ONS does not accept responsibility for any onward use of the output.