Academies in England are publicly funded independent schools with greater autonomy over curriculum design, school hours, term dates, and staff pay and conditions. The first academies appeared in England in 2002 during the Labour government (1997-2010) to address the underperformance of some urban schools. Schools that persistently underperform are required to convert into sponsored academies, gaining independence from Local Authority control and increasing their autonomy. However, it was not until 2010 with the Coalition Government (2010-2015) and the passing of the 2010 Academies Act that the number of academies increased significantly. This enabled the transformation of academies to any school without the need to have a sponsor. The study aims to examine, whether the increase in school autonomy by English schools due to the expansion of academies explains the widening academic performance gap between English and Welsh upper secondary students.

Numerous studies have explored the impact of school autonomy on student performance. For example, Hanushek et al. (2013), found substantial differences in the impact of decentralised decision-making depending on the economic and educational development of the country. Countries with stronger institutions benefit from increased autonomy of their schools, while in countries with lower levels of economic development, more school autonomy appears to damage student outcomes. The heterogeneous effect of school autonomy on academic performance is also found across different types of students within countries. Irmert et al. (2023) found little evidence that autonomous schools have an overall effect on achievement across 15 countries over 16 years. However, their results indicate that students from high socio-economic backgrounds benefit from autonomous schools, while immigrant students are disadvantaged.

Closely related to our study, the literature examines the expansion of academies in England, which can be grouped into two types: sponsored academies or converter academies. Machin and Vernoit (2011) and Eyles and Machin (2019) examine the impact of academy conversions between 2001/02 and 2008/09 on secondary school students’ achievement. They found an increase in achievement, which was more conclusive for schools that had been academies for a longer period and for those that experienced a greater increase in school autonomy. On the other hand, Machin and Silva (2013) found a positive impact of academies that converted between 2002 and 2007, mainly due to increased attainment among students in the top half of the ability distribution, especially those in the top 20% tail. However, they found little evidence that academies improved outcomes for students in the lower 10% to 20% of the ability distribution.

This strand of literature also provides evidence on the impact of conversion academies after 2010, particularly converter academies at the primary level. The main database used in these studies is the National Pupil Database (NPD), and they analyse the causal impact of academies on student performance using the difference-in-differences technique. McDool (2016) examined academies created during the academic years 11/12, 12/13, and 13/14. Her results indicate a positive and significant impact of conversion academies on students' average grades. The effect is more consistent for academies in less deprived neighbourhoods. Regarding student characteristics, she found that white students and students who are not eligible for free school meals consistently benefit from attending an academy. Regan-Stansfield (2018) looked at primary academies that converted between 2012/13 and 2014/15, whereas Eyles et al. (2017) looked at primary academies that converted between 2010/11 and 2014/15. Both found no significant average impact of academic performance. However, Regan-Stansfield (2018) found a small positive effect for students eligible for free school meals.

The present study uses the PISA database which tests 15-year-old students in reading, mathematics and science. Moreover, it provides information on the characteristics of students, their families, and their schools. The large UK sample across PISA editions allows us to compare England and Wales's results. Note that England scores higher than Wales in all three competences tested in PISA (reading, mathematics and science). This gap has widened over time. However, looking at performance levels, the gap is smaller for lower-performing students, yet widens significantly for higher-performing students, especially in mathematics and science. Moreover, we document a significant increase in the number of private government-dependent schools in England, mostly academies, while the share of public schools has been decreasing. In Wales, however, there has been little change in the distribution of schools by ownership. This indicates that England has more school autonomy than Wales which has increased in recent years. The empirical model we use to estimate the impact of increasing school autonomy on academic performance is a difference-in-differences (DID) estimator. Our dependent variable is the students’ scores in mathematics, reading and science. In our specification, the treatment group is composed of students from England who have experienced an increase in school autonomy whereas the control group is students from Wales. The pre-treatment period is PISA 2006, PISA 2009 and PISA 2012 and the post-treatment period is PISA 2015. To examine the heterogeneous effects that increased school autonomy may have on different student profiles, we use an extension of the DID estimator, called difference-in-difference-in-differences (DDD). We examine differences by gender (male and female students), immigrant status (natives and immigrants) and index of economic, social and cultural status. For the DID and DDD estimates, we have tested parallel trends of the treatment and control groups in the absence of treatment. To examine the impact of the treatment on different levels of academic performance we estimated a quantile regression.

The results indicate a positive impact of increased school autonomy in science. In addition, there are no significant differences observed between genders or immigrant status, with only marginal disparities by socio-economic level. Additionally, the analysis by achievement level reveals that, in science, all students demonstrate an improvement in their scores, except for those in the lowest achievement decile. However, the impact of school autonomy on reading and mathematics scores was not statistically significant on average. Note that only students from the fourth achievement decile onwards are favoured by the increase in school autonomy. The differences observed between the subjects can be explained by the challenges that the English education system has experienced in recent years. To illustrate, England has had considerable issues recruiting and retaining teachers where, some subjects suffer more from teacher shortages than others. Physics, computing, religion, design and technology, and modern foreign languages are among the subjects with the greatest difficulty in recruiting teachers. A shortage of teachers can result in teachers teaching specialist subjects without the relevant qualification, in the reduction of subject provision or in students not continuing with the affected subject at higher levels of education (House of Commons, 2024). Uncompetitive teacher salaries are one of the factors behind the shortage of teachers, particularly in science, mathematics and computing (SMC). Teachers in SMC subjects have better-paid options outside of teaching (Worth & Van den Brande, 2019). Academies have a greater degree of autonomy than maintained schools in making decisions about staffing and salaries. As a result, academies have more autonomy to attract teachers of science subjects - where there is a shortage of teachers- by offering them better salaries or working conditions. This would explain why the increase in the number of academies has had such a positive effect on science performance.

Eyles, A., & Machin, S. (2019). The introduction of academy schools to England's education. Journal of the European Economic Association, 17(4), 1107-1146. Eyles, A., Machin, S., & McNally, S. (2017). Unexpected school reform: Academisation of primary schools in England. Journal of Public Economics, 155, 108-121. Hanushek, E. A., Link, S., & Woessmann, L. (2013). Does school autonomy make sense everywhere? Panel estimates from PISA. Journal of Development Economics, 104, 212-232. House of Commons (2024). Teacher recruitment, training and retention. Second Report of Session 2023–24. Irmert, N., Bietenbeck, J., Mattisson, L., & Weinhardt, F. (2023). Autonomous Schools, Achievement, and Segregation. Machin, S., & Silva, O. (2013, March). School structure, school autonomy and the tail (CEP Special Paper No. 29). Retrieved from http://cep.lse.ac.uk/pubs/download/special/cepsp29.pdf. Machin, S., & Vernoit, J. (2011). Changing school autonomy: Academy schools and their introduction to England’s education (CEE Discussion Paper No. 123). Retrieved from http://cee.lse.ac.uk/ceedps/ceedp123.pdf. McDool, E. (2016). The effect of primary converter academies on pupil performance (Sheffield Economic Research Paper Series No. 2016013). Retrieved from https://www.sheffield.ac.uk/polopoly_fs/1.670238!/file/paper_2016013.pdf. Regan-Stansfield, J. (2018). Does greater primary school autonomy improve pupil attainment? Evidence from primary school converter academies in England. Economics of Education Review, 63, 167-179. Worth, J. and Van den Brande, J. (2019). Retaining Science, Mathematics and Computing Teachers. Slough: NFER.