Determinants of Human Capital Formation Inauguraldissertation zur Erlangung des Grades eines Doktors der Wirtschafts- und Gesellschaftswissenschaften durch die Rechts- und Staatswissenschaftliche Fakultät der Rheinischen Friedrich-Wilhelms-Universität Bonn vorgelegt von Renske Adriana Stans aus Vlaardingen, Niederlande 2020 Dekan: Prof. Jürgen von Hagen Erstreferent: Prof. Pia Pinger Zweitreferent: Prof. Thomas Dohmen Tag der mündlichen Prüfung: 21.
August 2020 Acknowledgments Roughly five years ago, I started the doctoral program at the BGSE with a mindset towards research that can be best described as uninformed optimism.1 In the years that followed, reality set in, projects did not always work out as expected, and I encountered occasions of pessimism and a few times even despair. Yet, these difficult moments were also the times in which I grew the most, both academically as well as personally. Moreover, the hard times have brought me the informed optimistic view on research that I have today, and for which I am grateful. Nevertheless, this all would not have been possible if it were not for the people who stood by me along the way.
I want to thank my supervisor Pia Pinger, for patiently answering my endless stream of questions and her valuable advice on research and academia. Pia, thank you for everything, I could not have wished for a better supervisor, and I hope we will meet often in the future. I also want to thank my second supervisor, Thomas Dohmen, for our productive discussions about research in truly its broadest sense. These meetings have often helped me to gain new insights, refresh old ones, or gave me new motivation.
Great thanks go to the entire faculty at the BGSE, IAME and briq for creating a productive research environment, providing academic encouragement, and ensuring financial support. In addition, I want to thank Silke Kinzig, Britta Altenburg, Andrea Reykers and Simone Jost for their administrative help all those years. I feel very fortunate for my fellow graduate students, in particular Lasse, Sven, Marina and Marius, for not only helping me to get through these last five years, but by making them fun as well. Above all I am thankful for Laura, having you by my side (both literally and figuratively) means more to me than you can ever imagine.
To my friends outside academia: Annet, Esmée, Jonas, Josephine, Kees, Lisanne, Luis, Marjolein and Sophie, thank you for being there when I needed you, providing me with welcome distractions, and giving me the energy to keep going. And last, but not least, I am immensely grateful for my parents Erna and Michel, my sister Annemarieke and brother-in-law Albert. Thank you for believing in me - especially at times that I did not - and for reminding me what really matters. 1 Common phases of a PhD are described in Julio Peironcely’s blog on NextScientist.com ii Contents List of Figures v List of Tables vii 1 Introduction 1 References.
4 2 The (Expected) Signaling Value of Higher Education 5 2.1 Perceived Wage Returns .2 Perceived Non-Wage Returns .3 Origins of Returns .4 Perceived Signaling Value of Higher Education .1 Immediate Wage Returns .2 Immediate Non-Wage Returns .3 Persistence of the Graduation Premium .5 Implications of the Signaling Theory .1 Heterogeneities in Signaling .2 Determinants of Leaving .A Additional Figures and Tables .B Counterfactual Labor Market Questions .C Data-Cleaning Rules. 56 CONTENTS 3 A Setback Set Right? Unfortunate Timing of Family Distress and Educational Outcomes 59 3.2 Background and Data .1 The Dutch Education System .2 Short-Term Effects .3 Long-Term Effects .A Additional Figures and Tables .B Bridge Class Ambiguity. 97 4 Parental Investments and Environmental Incentives 99 4.A Additional Figures and Tables. 119 iv List of Figures 2.1 Shares of expected wage trajectory patterns .2 Density of starting wages and returns .3 Density of job satisfaction and job-finding probability .4 Graduation premium and the development of university-leave wages .5 Expected yearly wage over the life time by work productivity .6 Coefficients of fixed effect model with semester dummies .A1 Computed parameters of the mincer wage equation .A2 Expected wage trajectory patterns .A3 Expected starting wage after graduating by gender and major .A4 Density of job finding probability at labor market entry .A5 Expected yearly wage over the life time, conditional on diminishing wage differences .1 Weekly frequency of grandparental death .2 Track placement test score by time of grandparental death .3 Density of number of correct answers by treatment status .4 Probability of attending the vocational track by time of grandparental death .A1 The Dutch education system .1 Yearly unemployment rate by federal state .2 Heterogeneous effects of regional unemployment on parental academic in- terest .3 Heterogeneous effects of regional unemployment on parental homework help110 4.A1 Heterogeneous effects of regional unemployment on hiring a tutor.
119 LIST OF FIGURES vi List of Tables 2.2 Immediate wage returns .3 Immediate non-wage returns .5 Heterogeneities in immediate wage returns .6 Determinants of probability to leave university .A1 Immediate wage returns (bachelor students) .A2 Immediate non-wage returns (bachelor students) .A3 Immediate wage returns with semester dummies .A5 Heterogeneities in immediate wage returns (bachelor students) .1 Percentage of children per track by grade .3 Effect of grandparental death on track placement test outcomes .4 Effect of grandparental death on makeup test participation .5 Effect of grandparental death on teacher advice .6 Effect of grandparental death on initial track placement .7 Effect of grandparental death on switching tracks .8 Effect of grandparental death on track attendance .9 Robustness analysis: treated control group .10 Robustness analysis: selection .11 Robustness analysis: time patterns .A3 Discrepancy teacher advice and standardized test performance by treat- ment status .A4 Effect of standardized test on initial track placement .A5 Effect of grandparental death on graduation outcomes .A6 Heterogeneous effects of grandparental death on track placement test out- comes. 94 LIST OF TABLES 3.A7 Heterogeneous effects of grandparental death on makeup test participa- tion, teacher recommendation and track switching .A8 Effect of grandparental death on track attendance excl.B1 Hypothetical percentages of children per track in grade 7 and 10 .1 The effect of regional unemployment on parental investments .2 Heterogeneous effects of regional unemployment on maternal support .3 Robustness of the effect of regional unemployment on parental investments113 4.4 The effect of regional unemployment and inequality on parental worries.A1 Survey items of parental investment variables .A2 The effect of regional unemployment and background characteristics on parental investments .A3 Heterogeneous effects of regional unemployment on paternal support and school contact .A4 Robustness of the effect of regional unemployment on parental investments123 4.A5 The effect of regional inequality on parental investments. 124 viii Chapter 1 Introduction Human capital is a key source of economic growth in current developed societies. More than 70 percent of the total amount of wealth in high-income OECD countries stems from human capital (Hamilton et al.
Likewise, cross-country differences in GDP growth can often be traced back to variation in cognitive skill levels (Hanushek and Woessmann, 2012). Also from an individual’s perspective accumulated human capital is important for economic prosperity. Key later-life outcomes such as labor earnings and probability to be unemployed are increasingly dependent on an individual’s human capital (Acemoglu and Autor, 2011). As such, differences in human capital are largely responsible for existing inequalities within society, making it essential to understand the determinants of individual human capital formation.
The foundations of economic research regarding human capital formation are made by the work of Becker (1962). Human capital can be seen as broad as an individual’s knowledge, skills, ideas, and health that improves the efficiency of the human factor, and Becker was one of the first to approach this as an economic concept. He saw human capital no different than any other type of capital, in terms of the investments that can be made and the returns it generates. His original human capital investment models focus on investments in education, which is one of the main contributors to human capital.
In these models individuals are assumed to invest in education until their marginal returns equal their marginal costs, and variation in investment levels mainly comes from individuals facing different financial constraints (Becker, 1994). Since the development of the original human capital investment models there has been an enormous increase of economic studies that focus on gaining a better under- standing of why individuals accumulate different levels of human capital. The prevalent explanations in the literature can roughly be divided along two dimensions. First, there may be obstacles or constraints on the investment side that hinder people to accumu- late human capital.
Second, variation in human capital formation may be explained by differences in returns between individuals, both in terms of actual and expected returns to obtaining human capital. INTRODUCTION This dissertation presents three empirical papers that explore various hypotheses related to both the investment and return dimension, of why there may be differences in individual human capital formation: Do people expect returns from accumulating human capital, and if so where do they belief these returns originate from? Can temporary events of distress in childhood have longstanding negative consequences on human capital formation in the presence of standardized tests? Do worsening local economic conditions incentivize parents to invest in children’s human capital? By answering these questions, this dissertation contributes to coming a step closer to identifying determinants of human capital formation. Chapter 2 provides new insights to the long-standing debate between human capital formation versus signaling as an explanation for returns from attending higher educa- tion. According to the signaling theory obtaining an educational degree does not form human capital, it merely reveals it, as only high-ability individuals can obtain the signal (Spence, 1973).
The paper is joint work with Laura Ehrmantraut and Pia Pinger, and its contribution is twofold: first, we estimate the perceived premium to obtaining higher education for university students; second, we investigate whether students ascribe the premium to acquired human capital or the signaling value of the degree. Accordingly, we conducted a survey among a large and diverse sample of German students at different stages of higher education to elicit counterfactual labor market expectations for the hy- pothetical scenarios of leaving university with or without a degree. These expectations are collected for the time when individuals start their first job and at age 40 and 55 to explore developments throughout the working life. Our findings indicate substantial perceived returns to finishing higher education, not only in terms of earnings but also with respect to job satisfaction and the probability of finding a suitable job.
To estimate the perceived importance of signaling in generating these returns, we employ a within-individual fixed effects model. This strategy circum- vents selection bias between university-leavers and university-graduates, as it compares the leaving and graduating scenario within individuals. We document that the perceived returns from signaling are substantial, as obtaining a degree raises returns by roughly 20 percent, whereas one more semester of accumulating human capital in university does not significantly raise returns. Moreover, the importance of signaling at the start of one’s career is expected to largely persist over an individual’s working life.
As the find- ings show that people expect an extensive part of the returns to come from signaling, differences in human capital formation may partially reflect that people have different inherent abilities and obtain different signals. Chapter 3 contributes to the extensive stream of literature that explains differences in human capital investments by the existence of various obstacles, such as constraints related to income, time, attention, or institutions. In this chapter I look at how the consequences of another obstacle, namely experiencing an event of temporary family distress, may be aggravated or diminished by prevalent features of education systems. In 2 particular, I investigate how children’s educational outcomes are affected by experiencing a common form of family distress - the death of a grandparent - shortly before taking a high-stakes standardized test.