research

Publications

Working Papers

  • College Majors and Earnings Growth (with Woosuk Choi, Josh Kinsler, and Ronni Pavan)
    • Abstract In this paper we estimate major specific earnings profiles using matched American Community Survey (ACS) and Longitudinal Employer-Household Dynamics (LEHD) data. The advantage of the matched data relative to the ACS alone is that it provides a long panel of worker earnings, thus avoiding estimating life cycle profiles using cross- cohort variation. Once we allow the returns to major to vary by cohort, we find that engineering, computer science, and business majors experience faster earnings growth relative to humanities majors. For example, the gap in earnings between technical majors like engineering and computer science and humanities grows by 5-6% between ages 23 and 50. Our estimates also indicate that more recent graduates in these fields earn a larger premium relative to humanities than earlier cohorts.
    [Manuscript]
     
  • Neighbors' Spillovers on High School Choice (with Juan Matta)
    • Abstract Do neighbors affect each others’ schooling choices? We exploit oversubscription lotteries in Chile’s centralized school admission system to identify the effect of close neighbors on application and enrollment decisions. A student is 7-10% more likely to rank a high school as their first preference and to attend that school if their closest neighbor attended it the prior year. These effects are stronger among applicant-neighbor pairs with lower education, college expectations, and prior academic achievement, measured by previous scores in national standardized tests. Lower-achieving applicants are more likely to follow neighbors to schools with better attributes when their closest neighbor’s test scores are higher. Our findings suggest the existence of frictions that prevent some families from learning about all available schools. Targeted policies aimed at increasing information to disadvantaged families have the potential to alleviate these frictions and generate significant multiplier effects.
    [Manuscript]
     
  • Licensure Tests and Teacher Supply (with Marcus Winters) - Submitted
    • Abstract We apply a sharp regression discontinuity design to administrative data from Connecticut to investigate the impact of failing the first attempt at a licensure test on teacher supply. We find deterrent effects from failing both a basic skills test required to enter an educator preparation program (Praxis I) and a subject-matter test used for ultimate certification (Praxis II). Failing Praxis II especially deters those seeking endorsement to teach within the shortage areas of STEM and special education. Failing Praxis I especially deters those who would be less effective teachers, but failing Praxis II disproportionately pushes out relatively effective potential teachers.
    [Manuscript]
     
  • Teacher Mis-assessments and High School Outcomes
    • Abstract Does mis-assessment by teachers on subjective evaluations matter for students’ educational outcomes? I employ administrative data from North Carolina that contain standardized test scores and teacher assessments for each ninth-grade student to examine whether exposure to a teacher whose judgments differ systematically from students’ achievement levels impacts student outcomes. Exposure to teachers who are more likely to overassess students, relative to what test scores signal, increases GPA and college expectations for girls and non-white students. In terms of SAT scores, I find increases for blacks and Hispanics but decreases for Asian students.
    [Manuscript]

Work in Progress

  • Aspirations, Education, and Occupations (with Kegon Tan)
    • Abstract It is well documented that children often “inherit” their parents’ occupations. This paper studies the role of early occupational aspirations in determining later life outcomes, a potentially important channel for intergenerational correlations in occupations. Using the Wisconsin Longitudinal Study, we estimate a lifecycle model of college choice and occupation choice to quantify the effect of aspirations on education and wages. We find that aspirations have a sizeable impact on educational attainment and wages, even conditional on latent skills that we recover from the choice model. We also simulate the importance of family background conditional on skills through the strong correlation between family background and aspirations. Our findings suggest that aspirations may be valuable for reducing intergenerational inequality.
     
  • Coworker Networks, Residential Neighbors, and Earnings Growth - W.E. Upjohn Institute Early Career Research Award
    • Abstract I study the relevance of former coworkers and neighbors to explain job mobility and earnings growth patterns for workers in the United States. Using matched employer-employee data from the Longitudinal Employer-Household Dynamics I construct individual-level proxies of new job opportunities available to workers using the number of new hires in firms where previous coworkers and close neighbors work. I extend the empirical framework of Caldwell and Harmon (2019) to isolate the effect of individual-level networks from other determinants of job mobility. I merge these records to multiple rounds of the American Community Survey to investigate heterogeneity across occupations and college majors.
     
  • College Major Choice and Labor Market Matching (with Choi, Kinsler, and Pavan)