This paper estimates the effect of school-entry age on grade repetition and academic achievement in a middle-income education system characterised by high repetition, institutional flexibility in Grade 1 admissions, and relatively strong compliance with official school-entry age requirements. South African school administrative records (SA-SAMS) form the basis of a learner-level panel dataset that is used to follow a cohort of first-time Grade 1 entrants from 2018 through to Grade 4.
Although legislation permits school-entry over an 18-month window, schools effectively implement one of two de facto regimes. Either they admit learners in an earlier “Mid-year” or later “Calendar Year” age range. Because parents may also select to send their children early or delay their entry based on unobserved readiness, the realised entry age is endogenous. To address this selection, actual school-entry age is instrumented with expected school-entry age, which is calculated based on learners’ birthdays and the school-entry age regime. School fixed effects account for the lack of standardisation in outcome measures across schools.
Relatively older entrants perform substantially better in the early grades. Instrumental-variables estimates indicate that being one year older at entry reduces the probability of repeating Grade 1 by roughly 16 percentage points, relative to a mean of 13% at Mid-year schools, and by 10 percentage points for a mean repetition rate of 10% at Calendar Year schools. Older learners also achieve higher marks in home language, first additional language, and mathematics. Relative age effects are larger for boys and for learners attending poorer schools. Although these gaps attenuate, meaningful differences persist to Grade 4. Non-parametric rank-mobility analyses show that convergence is higher among lower-ranked learners in Grade 1 and is correlated with higher rates of repetition.
DOWNLOAD PDF
