Do neighbours affect teenage outcomes?: evidence from neighbourhood changes in England

In this paper, we use census data on several cohorts of secondary school students in England matched to detailed information on place of residence to investigate the effect of neighbours’ background characteristics and prior achievements on teenagers’ educational and behavioural outcomes. Our analysis focuses on the age-11 to age-16 time-lapse, and uses variation in neighbourhood composition over this period that is driven by residential mobility. Exploiting the longitudinal nature and detail of our data, we are able to control for pupil unobserved characteristics, neighbourhood fixed-effects and time-trends, school-by-cohort unobservables, as well as students’ observable attributes and prior attainments. Our results provide little evidence that neighbours’ characteristics significantly affect pupil test score progression during secondary education. Similarly, we find that neighbourhood composition only exerts a small effect on pupil behavioural outcomes, such as general attitudes towards schooling, substance use and anti-social behaviour.