Internal Lunch Seminars

This paper explores the success of scaling up of state programs. I study the two phases of staggered rollout of the Advanced Placement (AP) Incentives Programs in Texas which aimed to increase AP utilization. The 1997 pilot phase included a small group of high schools, the Scaled-up phase expanded this to half of all Texas high schools in 2001. Using a staggered difference-in-differences approach, I separately estimate the impact of the program in each phase. I find a 70% increase in AP Enrollment and a 40% increase in the number of AP courses for the schools which implemented the program before scaling up. College enrollment and graduation increases by 10% and 3%, as well as a 7% increase in wages. However, there is a null effect for the outcomes of the schools which implemented the program after scaling up. The potential explanation for the disparate effects is that the student demand for the AP courses differed between two treatment arms.

I use deviations from the expected age at menarche to estimate the marginal treatment effects (MTEs) of teen childbearing on schooling and labor outcomes for Ecuadorian mothers and schooling and health outcomes for their firstborn children. Findings suggest that women with unobservable characteristics that make them less likely to become teen mothers are less likely to participate in the labor force, have fewer years of schooling, and are less likely to finish high school if they become teen mothers. Women with values of unobservables that make them more likely to become teen mothers do not have their schooling attainment negatively impacted and increase their labor force participation. I do not find evidence of effects on firstborn children. These findings help reconcile seemingly conflicting evidence from past studies and imply that there is potential to improve women's outcomes by reducing teen childbearing rates when opportunity costs are sufficiently high. However, these findings counter the belief that teen childbearing has been a significant source of intergenerational transmission of low socioeconomic status.

In many contests, players are not aware of how many competitors they face. While existing studies examine how disclosing this number affects their productive effort, this paper is the first to consider its impact on destructive behavior. For doing so, I theoretically and experimentally study how revealing the number of contestants affects both effort and sabotage compared to concealing this information. Further, I evaluate the created value by comparing the resulting performances, which are shaped by the combination of the exerted effort and the received sabotage. I show that the overall performance can be higher under concealment, even though the disclosure policy does not affect average effort and sabotage levels. The experimental results largely confirm these theoretical predictions and demonstrate the significance of accounting for the effects of sabotage, as it induces performance differences between the group size disclosure policies. By concealing the number of contestants, a designer can mitigate the welfare-destroying effects of sabotage, without curbing the provision of value-creating effort.