ABSTRACT
The debate regarding the relative effects of automatic grade promotion versus grade retention has long concerned education policymakers and researchers. This work applies a random-effects probit model to assess the probability of early dropout in Brazilian public schools based on a unique administrative database which permits to infer household income at the student level for more than 60 thousand pupils enrolled in 273 schools. The results suggest evidence that decisions to drop out early are significantly influenced by whether approval has taken place and age-grade distortion. Moreover, simulation exercises have suggested that automatic grade promotion policy could reduce up to 35% the probability of a low-income student dropout from secondary school, while 24% for their rich peers.
Disclosure statement
No potential conflict of interest was reported by the authors.