Impacts of delayed school entry on child learning in rural northwestern China - forced delay versus voluntary delay

Most previous studies exploit the discontinuity in children's school entry age around the enrolment cut-off to identify the effect of delayed school entry on child learning. However, the effect so identified is only relevant to children who were 'forced' to enter school late because of being born slightly after the enrolment cut-off. In developing countries, many children voluntarily start school late regardless of their birthdates, for whom the commonly-used discontinuity-based strategy fails to identify the effect relevant to their delayed enrolment. This study exploits community peer effects to estimate the impact of voluntary delay in school entry, using older peers' school entry age to instrument one's own school entry age. Analysing data on nearly 4,000 children from rural northwestern China, we found that while both types of delay in school entry raise the incidence of first-grade repetition, the effect of voluntary delay is much larger than that of forced delay. More specifically, one year of voluntary delay increases a child's likelihood of first-grade retention by 12.4 percentage points, which is more than twice the effect of forced delay.