Human Capital Returns to Education: Three Essays on the Causal Effects of Schooling on Skills and Health (Dissertation)

Abstract

This dissertation estimates returns to education in terms of skills and health as important aspects of human capital given their importance in determining economic outcomes. The dissertation is based on three independent empirical research articles which constitute the three main chapters 2, 3, and 4, and a comprehensive introduction in Chapter 1 and conclusion in Chapter 5. The analyses exploit several exogenous variations in German educational policies and apply state-of-the art microeconometric techniques for causal inference. The chapters complement each other in three dimensions: First, each chapter looks at different aspects of education, together considering both qualitative measures - including the curriculum, learning intensity, and timing - as well as quantitative factors - like years of schooling. Second, while all chapters estimate non-monetary effects of education, each investigates a specific dimension of human capital, overall addressing cognitive skills, non-cognitive skills, and health. Third, to complement each other from a life-course perspective, different age groups are analyzed including seventeen-year old adolescents still enrolled in secondary education, young adults at the time of graduating from high school, and adults between the ages 50 and 85, long after the completion of secondary education. Chapter 2 investigates the short-term effects of a reduction in the length of high school on students' personality traits using a school reform carried out at the state level in Germany as a quasi-natural experiment. Starting in 2001, academic-track high school education was reduced by one year in most of Germany's federal states, with the overall curriculum left unchanged. This enabled students to obtain their university entrance qualification after only 12 rather than 13 years of schooling, but it increased learning intensity through an increase in weekly class hours. Exploiting the variation in the length of high school over time and across states, the effect of learning intensity on students' Big Five personality traits and on their locus of control is identified. Using representative data on adolescents and young adults from the German Socio-Economic Panel (SOEP) study, the estimates show that shortening high school caused students to be less emotionally stable. Furthermore, the results point to important heterogeneous effects. In addition to regional and gender differences, students not living with both parents and students with migration background showed stronger personality changes following the reform: they became more open and extroverted, and more open and conscientious, respectively. Thus, the chapter concludes that the educational system plays an important role in shaping adolescents' personality traits. Chapter 3 focuses on cognitive skills and investigates two mechanisms through which they may be affected by education. I rely on the same reform analyzed in Chapter 2 for causal identification but conduct two separate analyses based on different datasets: First, I exploit the variation over time and across states to identify the effect of an increase in class hours on same-aged students' intelligence scores, using SOEP data on seventeen year-olds. Second, I investigate the influence of earlier instruction at younger ages, using data from the German National Educational Panel Study (NEPS) on high school graduates' competences. The results suggest that overall, secondary education impacts students' crystallized cognitive skills in adolescence especially through instructional time rather than through age-distinct timing of instruction. However, they also reveal that increasing instructional time aggravates gender differences in numeracy. Chapter 4 analyzes whether education has a protective effect on mental health. To estimate causal effects, an instrumental variable (IV) technique is implemented with two different instruments to estimate local average treatment effects (LATE) at different parts of the educational distribution: (i) a reform extending compulsory schooling by one year implemented in West German federal states between 1949 and 1969 and (ii) the individual availability of higher education measured by the spatial distance to the nearest university at age 19. The analyses are based on rich individual SOEP data on adults aged 50 to 85, augmented by detailed information on universities from the German Rectors' Conference. Analyses on the Mental Component Summary (MCS) score as a generic measure of overall mental health are complemented by disorder-specific diagnoses. Results support existing evidence on a positive relationship between completed years of secondary schooling and mental health in standard Ordinary Least Squares (OLS) estimations. However, the IV estimations reveal no such causal protective effect. If any, the estimates point towards a negative effect among the lower educated. These results are confirmed when explicitly modeling effect heterogeneity through marginal treatment effects.