This paper examines the causal link between educational success and teenage childbearing. Previous studies show that girls are less likely to become teenage mothers if they have to complete more years of compulsory schooling (Black, Devereux and Salvanes 2008; Silles 2011), suggesting that education has a negative impact on teenage childbearing. On the other hand, McCrary and Royer (2011) finds little evidence that education has any impact on teenage fertility behavior. The authors use a regression discontinuity framework to compare girls who are born right before the school year admission cutoff date (who can start school with the rest of their birth cohort) and girls born right after the cutoff date (who must start school one year later). They find that girls who are born right after the cutoff date complete fewer years of education, but are no more likely to have teenage births. They conclude that education has little impact on fertility behavior.

This paper follows a similar empirical strategy as McCrary and Royer (2011) but uses high quality restricted-access data obtained from the North Carolina Education Research Data Center. While the dataset used in McCrary and Royer (2011) only has information on individuals’ self-reported age and education at the time of birth, this dataset provides administrative data on their actual age at the time of birth, as well as their complete educational records up to age 21, including their scores on end-of-grades tests in the third and eighth grades. The difference in data quality makes a large difference to the analysis. Using this more detailed dataset, Cook and Kang (2013) show that while girls who are born right after the cutoff date (and start school one year later) complete fewer years of education, they also have much higher test scores. Hence, the impact of being born after the cutoff date on educational success is much less negative than implied in McCrary and Royer (2011), which may explain why these girls were not more likely to become teenage mothers. This essay formally tests this hypothesis by using an instrumental variable regression model to distinguish the impacts of years of education and test scores on teenage childbearing. My preliminary findings suggest that, contrary to McCrary and Royer (2011), years of education and test scores both have a negative impact on teenage childbearing. To my knowledge, this essay is the first study which takes into account performance on school tests while using school entry policies to examine the impact of female education on teenage childbearing, as well as the first study which distinguishes the impacts of years of education and test scores.