The demand and supply of public education

Education plays a central role in securing peoples' well-being and has important social returns. Numerous studies have documented the relationship between education and economic growth, and economic and social inequality. The increased importance of schooling is mirrored by congruent increases in investments in terms of money and time. More people spend more time in the education system, and these expansions in enrollment have led to large increases in total spending on education. The provision of education is a major public sector activity around the world. This large public involvement in the financing and provision of education raises the question of how to organize schools and how to organize the schooling sector. There are many evaluations of specific educational treatments such as going to high-school, completing a vocational degree, studying business, etc depending on the study at hand. While this type of study tells us about the effects for those who were treated relative to a specific alternative treatment, they do not directly answer the question who should be treated, or what treatment among a broader choice is better. People are often constrained in their choices. This happens when there are limits on how many people can enroll in specific schools or programs. Such constrained schooling opportunities are often allocated based on how well students performed in school, or priorities derived from student characteristics such as location. One important challenge in setting and managing these constraints in the schooling sector concerns the optimal matching of students to slots. How do we get the right student in the right chair? The aim of the project is thus to better understand schooling choices and their consequences, and to study supply side policies. The first part of the project takes supply as given and studies the demand for schooling. Marriage market considerations are a classical demand factor.Our study of ethnic segregation in secondary education in Amsterdam finds that 40% can be attributed to demand and 42% to ability tracking, highlighting that segregation-reducing policies will involve less-preferred school assignments for many students. Demand factors can also lie beyond schooling and marriage market considerations are a classical example. We find that colleges are local marriage markets, mattering greatly for whom one marries, not because of the pre-determined traits of the admitted students but as a direct result of attending a particular institution at a given time. A new study of ethnic segregation in secondary education in Amsterdam finds that 40% can be attributed to preferences and 42% to ability tracking, highlighting that segregation-reducing policies will involve less-preferred school assignments for many students. In a related project we estimate and compare the earnings payoffs to post-secondary fields of study in Norway and Denmark. We find strong cross-country correlation in the payoffs to fields of study, especially after addressing methodological concerns that arise when there are multiple treatments. The second part of the project asks how to effectively target interventions and organize supply in a public school system. A new working paper investigates substitution effects in college admissions and shows that local expansions often have large displacement externalities elsewhere in the system. Ignoring these spillovers leaves 40% of the benefits of long-cycle STEM expansions unaccounted for. In a new ongoing project we investigate the complementary question of how the modality of admissions affects college outcomes. We also investigated the complementary question of how colleges can find successful applicants. GPA and alternative criteria such as interviews, essays, and tests provide information about candidates, but which work and why? We find that admission based on alternative criteria outperforms standard admission based on GPA. Alternative criteria are more effective in identifying good program matches, which ultimately leads to higher college completion rates. Most of the impact of alternative evaluation is found to be due to their impact on the applicant pool (sorting), and not because of they are better at identifying successful students keeping the applicant pool fixed (screening). The use of subject-specific grades (instead of GPA) leads to the admission of applicants that are less likely to succeed, while essays is the only instrument that is intrinsically better at screening. The use of tests, interviews and CVs do not outperform GPA in screening once we keep the applicant pool fixed. There is no evidence that interviews are an effective admission tool.

The large public involvement in the financing and provision of education raises the question how to organize the production and delivery of schooling. There is a vast experimental and quasi-experimental literature that evaluates the effectiveness of educational interventions and schooling. While these evaluation studies are very successful at recovering average causal effects for specific populations, it is less clear how to use these studies to extrapolate to other populations or treatment assignment mechanisms. So while we know the effects for those who were treated, this does not directly answer the question who should be treated. The aim of the project is to i) better understand schooling choices and their consequences, and ii) to bridge the ex-post evaluation literature and the ex-ante evaluation/treatment choice literature. To do so we will build on the Norwegian registry data, combine these data with targeted surveys, combine experimental and non-experimental methods, and apply frontier statistical and econometric methods. In the first part of the project we study the role of information, families and peers in preference formation how these affect education choices. We subsequently estimate the causal impact of these choices on educational trajectories, human capital accumulation and labor market outcomes. The second part of the project builds on these elements to study the supply side of education. This includes methodological work that is necessary to extrapolate findings to other populations and policies. The public provision of schooling involves equity and efficiency tradeoffs. The systematic investigation of treatment assignment from an explicit welfare point-of-view in this project makes these tradeoffs explicit. A final important contribution of the project is the quantification of these trade-offs for current and counterfactual policies.