Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Gender parity in education—an important global development goal—has been primarily measured through school enrollment, and the gender parity in education quality has received limited attention until recently. We address this issue by highlighting the intrahousehold allocation of education expenditure. We extend the hurdle model into a three-part model to enable decomposition of households’ education decisions into enrollment, total education expenditure, and share of the total education expenditure on the core component, or items relating to the quality of education such as private tutoring. We apply this model to four rounds of nationally representative household surveys from Bangladesh, a country that offers a unique setting in South Asia with the Female Stipend Programs (FSPs), a nationwide gender-targeted conditional cash transfer program. We demonstrate a strong profemale bias in the enrollment decision but contrasting promale bias in the other two decisions, conditional on enrollment. We argue that this contradirectional gender gap is unique to Bangladesh and that it can be explained partly by the FSPs. Both the three-part model and a separate analysis of double-difference model show that the FSPs promoted girls’ secondary school enrollment. However, the FSPs did not narrow the gender gap in the intrahousehold allocation of educational resources. Consistently, we find a gender gap in on-time completion of secondary school. Our findings collectively highlight the complex interplay of intrahousehold decisions and underscore the importance of minding the gender gap in the quality of education and implementing complementary policies to address it in developing countries.