Working Papers
Student Sorting and Academic Performance (Job Market Paper) |
|
This paper studies the effects of student sorting along ability on educational achievement. Using unique features of the Romanian high school admission process, I exploit the fact that differences in the number of high schools across towns lead to local variation in student sorting patterns. I first show that student sorting along ability is more prevalent in towns with more high schools. High-ability students attend schools with more positively selected peers if they live in places with more high schools. The converse is true for low-ability students. I exploit this finding to construct an instrument for peer quality by interacting the number of HS with student ability. Using this instrumental variable, I find that attending a high school with a one percentile higher average entrance grade improves a student's graduation score by 0.16 percentiles. Overall, sorting exacerbates educational outcome inequalities between high and low ability students in urban areas and inequalities between top students in urban areas and their rural counterparts. Focusing on school openings in a difference-in-difference-in-differences framework yields similar results. When a new school opens, sorting increases markedly and the achievement gap increases. Lastly, by matching novel data on school expenditures and teacher evaluations and mobility to student admissions and graduation records, I decompose the effects of attending a better school into three channels. Differences in teacher ability (48%), peer effects (25%) and school spending (5%) explain most of the effect of attending a better school. The results suggest that reallocating high ability teachers to underprivileged schools can alleviate inequalities in educational outcomes caused by student sorting.
|
The Effect of High School Duration on Educational Attainment and Labor Market Outcomes: Evidence from Ontario (coauthored with Xian Zhang)
Revise & Resubmit @ Canadian Journal of Economics, 2020
|
This paper studies a policy change in Ontario, Canada, which reduced the high-school duration from five to four years. By comparing the difference between pre- and post-reform cohorts in Ontario with those in the rest of Canada, we show the policy boosted the enrollment into post-secondary education while adversely affecting high school graduation and wage rate among post-secondary graduates. Two mechanisms were proposed to explain these findings: orientation effect and performance effect. We construct a dynamic discrete choice model to assess their relative importance, finding that orientation effect accounts for 11-20% of the decrease in the wage premia across educational categories.
|
Migration and Student Sorting
|
Using geocoded high school admission and graduation records from Romania, coupled with GIS data on highway construction, I study how transportation costs impact student rural to urban migration and how, in turn, this migration affects educational outcomes. In particular, I focus on the opening of highway segments in Romania and its impact on competition in high schools. I find that high-ability students, who are more likely to benefit from attending more selective urban high schools are more likely to migrate from rural to urban areas. This migration creates a crowding-out effect in urban areas, which harms lower-ability urban students and may harm low-ability rural students by decreasing their peers' quality. Rural to urban student migration responds to reduced transportation costs. This is particularly true for the migration of high-ability students.
|
Ongoing and Future Projects
|
Other
How Cooperative is Cooperative Federalism?: A Game of Legislative Supply Under Dualist and Double Aspect Federalism (coauthored with Christa Scholtz)
|
We explore the idea that cooperative federalism in Canada can actually lead to a decrease in cooperation between the federal and provincial governments when the governments act strategically. We build and solve a simple sequential game to describe the circumstances under which dual aspect federalism leads to more - and less - cooperation.
|
The Welfare Effects of Online Reviews: A Model of Observational Learning
|
This paper proposes a stylized model of observational learning and herding which describes how rational consumers make purchasing decisions based on online reviews. Consumers arrive in an exogenous sequence, draw a private signal and observed reviews posted by past consumers. Based on this information, they update their belief of the quality of the product and the utility they will derive from it. Conditional on purchase, consumers post reviews which can be consulted by subsequent consumers. The model predicts that the online review is welfare improving to consumers as it helps consumers learn product quality. However, this is not necessarily the case in the short run for low quality products, due to self-selection: higher signal consumers buy the product and post relatively high reviews, while lower signal consumers opt out. This self-selection can make estimating product quality difficult in the short-run. Moreover, self-selection also means that the difficulty of learning the product quality increases as the outside option improves. Unintuitively, this means that short-run consumer welfare can decrease with increased competition. Using review data from Amazon.com and a regression discontinuity design, reduced-form regression confirms that reviews are a driver of sales. Lastly, I estimate our model via a maximum likelihood approach for a number of Amazon.com products and make predictions of consumer welfare from reviews.
|
