Attendance of pre-primary institutions is positively associated with PISA test scores in most countries. Several indicators of structural quality of pre-primary education are identified that can account for the variation in the estimated coefficients on pre-primary attendance across countries, based on a cross-country student-level specification with country fixed effects and interactions between individual pre-primary attendance and country-level quality indicators. The association of pre-primary attendance with test scores at age 15 is larger in countries with higher per-pupil spending in pre-primary education, larger shares of children attending privately managed pre-primary institutions, and higher relative pay and higher levels of training of pre-primary teachers.