Research

Preference Estimation in Deferred Acceptance with Partial School Rankings (Job Market Paper), Preprint at  arXiv 2010.15960 [PDF]

The deferred Acceptance algorithm is a popular school allocation mechanism due to its strategy proofness. However, with application costs, strategy proofness fails, leading to an identification problem. In this paper, I address this identification problem by developing a new Threshold Rank model that models the entire rank order list as a one-step utility maximization problem. I apply this model to study student assignments in Chile. There are three critical contributions of the paper. I develop a recursive algorithm to compute the likelihood of my one-step decision model. Partial identification is addressed by incorporating the outside value and the expected probability of admission into a linear cost framework. An empirical application to Chilean data reveals that although school proximity is a vital variable in school choice, student ability is critical for ranking high academic score schools. Simulation exercises suggest that policy interventions such as tutoring aimed at improving student ability increase the representation of low-income low-ability students in better quality schools.

Featured in World Bank Development Impact Blog.

Centralized Admission Systems and School Segregation: Evidence from a National Reform, with Macarena Kutscher and Sergio Urzua, IZA Discussion Paper No. 13305, Submitted [PDF][Web Appendix]

This paper investigates whether centralized admissions systems can alter school segregation. We take advantage of the largest school-admission reform implemented to date: Chile's SAS, which in 2016 replaced the country's decentralized system with a Deferred Acceptance algorithm. We exploit its incremental implementation and employ a Difference-in-Difference design. Using rich administrative student-level records, we find the effect of SAS critically depends on pre-existing levels of residential segregation and local school supply. For instance, districts with prominent provision of private education experience an uptick in school segregation due to SAS. Migration of high-SES students to private schools emerges as a key driver.

Explaining Third Birth Patterns in India: Causal Effect of Sibling Sex Composition, working paper [PDF]

Son preference is a well-established phenomenon for India. This preference gets reflected in multiple dimensions of the childbearing process such as the size of the family and birth spacing. I use the sibling sex composition of the first two children to capture its impact on the third birth interval, induced by a preference for a son. Sibling sex composition provides a credible source of exogenous variation in the Indian context for births before 1990 as gender screening became widespread only after the economic reforms in 1990. My analysis shows that on average families with two sons face a 9% lower hazard of third birth relative to families with two daughters. This hazard ratio translates into a gap of roughly one month in the average third birth interval. I also show that sibling composition affects the proportion of third births spaced below 18 months, a critical cut-off for neonatal, post-neonatal and child mortality. Inter-birth intervals of less than 18 months increase the chances of the third child's mortality by 10% in my sample. A back of the envelope calculation based on these estimates suggest that about 1,500 infant births every year in India can be attributed to a higher proportion of daughters among the first two births.

When Is Not Safe to Go Outside: Implications of Rising Conflict on Female Physical Mobility in Afghanistan, with Cesar Cancho and Christina Wieser

Restrictions on women's physical mobility can impact human capital accumulation adversely, with enduring negative consequences for the present and the future. In this paper, we examine how the deteriorating security situation in Afghanistan in the early 2010s affected women's ability to mobilize out of their places of residence. We use the nationally-representative household budget surveys for 2007-2008 and 2013-2014, exploiting, for identification, the inter-temporal and geographic variation of each survey and conflict intensity. Our analysis shows that female mobility decreases by 4% relative to mean mobility if reported incidents increase by 10 units. This effect is sizable for Afghanistan as 25% of female respondents resided in districts that reported higher than 10 units increase in monthly incidents. We also found that educated women face lower mobility constraints, and this mobility premium for education rises with an increase in conflict intensity. Similarly, single women face fewer such constraints, more so in the presence of high conflict. This suggests that any peace agreement leading to violence reduction in Afghanistan can have substantial benefits for women, by lifting constraints to the use of services that contribute to human capital accumulation. Conversely, any policy intervention aimed at improving female human capital outcomes needs to ensure better mobility as a precursor.

Causal Inference for Deferred Acceptance Algorithms, with Guido Kuersteiner and Sergio Urzua, work in progress

Impact of Classroom Networks on High School Outcomes in Chile, work in progress