Causal Inference in Economics

In this meeting we will discuss and learn more about methods for causal inference popular in applied economics, touching on concepts like instrumental variables, difference-in-difference designs and synthetic control analysis.

Our keynote speaker is Wolter Hassink, professor of applied econometrics at UU, and newly appointed Dean of Research and Social Impact, Faculty of Law, Economics and Governance. Wolter will present a bird’s eye view on causal research designs in applied economics, with examples of his own work applying these methods in different areas (full abstract below).

We will have plenty of time for discussion and refreshments will be provided! If you want to be added to the Causal Data Science SIG mailing list and Teams environment, please contact o.ryan@umcutrecht.nl.

Keynote

Title: A bird’s eye view on causal research designs in applied economic research. Lessons from previous empirical work on refugees, sickness absenteeism, commutes and loneliness

Speaker: Wolter Hassink (Utrecht University School of Economics)

Abstract

First, in joint work with Emre Akgündüz and Marcel van den Berg (World Bank Economic Review, 2018), we investigate how the Syrian refugee inflows into Turkey affected firm entry and performance. To estimate the causal effects, we used instrumental variables, difference-in-differences, and synthetic control methodologies.

Second, I discuss the results of a paper jointly with Roberto M. Fernandez (The Manchester School, 2018), in which we exploit a natural experiment design of a firm relocation from Milwaukees Central Business District to the areas suburban ring in 1992. There is an exogenous source of variation on the adjusted commuting distance among those who stay at the firm. The estimates suggest that low-morale workers are responsive to the shock in commuting time for some of the dimensions of morale.

Third, Jordy Meekes and I investigated (Journal of Urban Economics, 2022) whether women and men cope with job loss differently. Using a quasi-experimental empirical design involving job displacement because of firm bankruptcy, we find that displaced women are more likely than displaced men to find a flexible job with limited working hours and short commutes.

Fourth, in ongoing research with Reneé van Eyden, we employ panel data of adult individuals from the South Africa. We estimate various reduced-form equations that measure the effect of loneliness on binary measures of the economic outcome variables employment, health and living conditions. To address the effect of simultaneity bias, we instrument the individual’s loneliness by instrumental variables that measure the person’s trustworthiness.