Individual Submission Summary
Share...

Direct link:

Ethnic Diversity and Well-being: Evidence from a Hundred-Thousand-Person Panel

Sat, October 2, 12:00 to 1:30pm PDT (12:00 to 1:30pm PDT), TBA

Abstract

De-segregation is often perceived to be a politically viable tool to improve integration and intergroup relations. Yet, humans have a strong preference for associating and bonding with similar others. Though ethnic mixing may advance integration and diminish prejudice, being surrounded by people who are ethnically dissimilar from oneself may come to the detriment of another essential societal factor: citizens’ well-being. Well-being is important not only because it improves a range of desired outcomes such as health and job performance; a society in which citizens thrive is arguably important in itself.

Yet, only few studies from different disciplines have investigated the link between ethnic composition and well-being and the findings are mixed. We argue that three main challenges remain that may explain these mixed findings. First, it is essential to measure ethnic composition at quite disaggregate levels to test whether ethnic interaction is a driver of well-being. Measures at the country, municipality, or even county level is often ill equipped to capture citizens’ experience with ethnic others in their daily lives. Second, it is central to focus both on the native as well as the minority population while taking into account that the minority population is not necessarily one coherent group. Third, it is important to improve causal identification. Former studies have addressed some of these challenges but not all of them at the same time. We aim to do so in this paper. Our main argument is that ethnic similarity fosters well-being and that this is a universal psychological phenomenon not bound to any specific ethnic group.

To test this hypothesis while meeting the above challenges, we use the rich Danish register data of the full population of Danish students in public schools from the fourth to the ninth grade (9-16 years old) from 2015 to 2018. Specifically, we couple information on Danish students’ ethnic background and schoolclass attendance with responses on a mandatory Danish National Well-being Survey. We use seven items from the Well-being survey that tap students’ school-satisfaction (e.g., “Are you happy about your school?”) and their positive relations with others (e.g., “Do you feel lonely?”) and create a single, reliable measure of well-being (alpha=0.86). Our final sample consist of ~700.000 observations.

This combination of register data and individual-level panel data allows us to more accurately and convincingly test whether ethnic composition affects well-being. First, it enables us to measure the (change in) ethnic composition objectively and precisely using a variety of operationalizations at a very low level of aggregation: the classroom. Second, having data on the full population allows us to estimate the effect for ethnic minorities efficiently. Third, the panel structure and the register data allow us to minimize selection. Besides from including a host of objective covariates at the student, parent, and classroom level as well as cohort and year fixed effects, we estimate the effect of ethnic composition on well-being using three models more suited for causal interpretation: school fixed effects, student fixed effects, and naturally occurring exogeneous variation in school changes. This enables us to jointly account for a host of unobserved factors by utilizing the within-school-between-class variation; the within-student variation; and exogenous changes over time that are less related to well-being and classroom ethnic composition.

We find universal support for the ethnic similarity hypothesis. As the share of non-natives in the classroom rises, the well-being of natives decreases and the well-being of non-natives increases. The effect sizes are moderate in size. These results remain the same across all three identifying models – school FE, student FE, and exogenous variation – as well as across different measures of the ethnic composition: the share of non-western immigrants, ethnic fragmentation, ethnic polarization, and the share of co-ethnics defined by country of origin. Importantly, however, minority citizens’ well-being depend more on the presence of other co-ethnics than by the presence of other non-natives broadly understood. This not only suggests strong support for the universality of the relationship; it also indicates that it may be theoretically misleading and methodolgocially imprecise to assume a general minority identity when studying well-being. Lastly, the presence of other co-ethnics has twice as strong an impact on minority citizens’ well-being compared to the well-being of native citizens.

These results suggest, first, that ethnic similarity breeds well-being. This seems to be a universal psychological phenomenon that operates across different ethnicities and dominance positions in society. Second, it indicates that de-segregation policies face a potential trade-off between improving societal integration on one side and the well-being of citizens on the other.

Authors