Complementarities in High School and College Investments

Abstract

How do pre-college investments affect college enrollment, graduation, and labor market outcomes? Using Swedish registry data, we document how students sort on multidimensional abilities into high school tracks, and sort on both abilities and high school track into college majors. These sorting patterns highlight the presence of dynamic selection on both the level and type of education, and suggest that there may be dynamic complementarities between secondary and post-secondary investments. With the goal of estimating dynamic complementarities, we develop a dynamic Roy model to account for selection and differential returns to abilities, prior investments, and other persistent unobservables. We identify the model using noisy measures of abilities combined with quasi-experimental variation at the high school and college application stages. We find heterogeneous returns to abilities by final education, and economically important dynamic complementarities between high school and college investments. Consequently, we find larger returns to counterfactual policies promoting investments in STEM skills in high school compared to in college.

Keywords: Cognitive and Non-cognitive Abilities, Education Investment, High School Curriculum, College Major Choice

JEL Classification: C32, C38, I21, J24, J31