Research

More than a century has passed since the abrupt exodus of 1.2 million Greek Orthodox from Anatolia and their resettlement in Greece, a transformative event for the country's social and demographic landscape. Today, more than one in three Greeks reports a refugee background. While its historical significance is well-documented, its short-, medium-, and long-term impact on human capital accumulation remains unexplored. How did forced displacement shape the educational trajectories of the uprooted and their offspring? Did refugees invest in portable skills to respond to uncertainty, or did they struggle to catch up with the autochthonous? To address these questions, we trace the educational investments of refugees and their descendants over the last 100 years, leveraging granular census data and a comprehensive mapping of both their origins in Anatolia and their settlements in Greece. The analysis provides compelling support for the uprootedness hypothesis. Though initially lagging, refugees settling in the Greek countryside eventually outperformed nearby natives in educational attainment. Their university choices also diverged, with refugees' lineages favoring degrees transferable beyond the Greek labor market, such as engineering and medicine, and natives specializing in law and other fields with a strong home bias. Exploring additional mechanisms reveals the critical role of linguistic barriers and local economic conditions in shaping these outcomes, rather than the refugees' pre-migration economic background. The widespread educational gains of refugees and their descendants over four generations offer some hope that the ongoing surge of forced displacement, despite its tragedy, if properly addressed by the international community, can be a backbone of economic resilience for the affected communities.

Analyses of spatial or network data are now very common. Nevertheless, statistical inference is challenging since unobserved heterogeneity can be correlated across neighboring observational units. We develop an estimator for the variance-covariance matrix (VCV) of OLS and 2SLS that allows for arbitrary dependence of the errors across observations in space or network structure and across time periods. As a proof of concept, we conduct Monte Carlo simulations in a geospatial setting based on U.S. metropolitan areas. Tests based on our estimator of the VCV asymptotically correctly reject the null hypothesis, whereas conventional inference methods, e.g., those without clusters or with clusters based on administrative units, reject the null hypothesis too often. We also provide simulations in a network setting based on the IDEAS structure of coauthorship and real-life data on scientific performance. The Monte Carlo results again show that our estimator yields inference at the correct significance level even in moderately sized samples and that it dominates other commonly used approaches to inference in networks. We provide guidance to the applied researcher with respect to (i) whether or not to include potentially correlated regressors and (ii) the choice of cluster bandwidth. Finally, we provide a companion statistical package (acreg) enabling users to adjust the OLS and 2SLS coefficient’s standard errors to account for arbitrary dependence. 

We investigate the long-term consequences of mass refugee inflow on economic development. After the Greco-Turkish war of 1919–1922, 1.2 million Greek Orthodox were forcibly resettled from Turkey to Greece, increasing the host population by more than 20% within a few months. To examine the long-term effects of this event, we build a novel geocoded dataset locating refugee settlements across the universe of more than four thousand Greek municipalities that existed in 1920. Using a battery of empirical strategies relying on different margins of spatial and temporal variation in the refugee inflow, we find that localities with a greater share of Greek refugees in 1923 display higher level of prosperity and industrialization sixty years after the event. These long-run benefits of refugees appear to be driven by the provision of new agricultural know-how and the transfer of technological knowledge in textile, which fostered growth through higher diversity in complementary skills. The economic gains of the resettlement were lower in places where refugees were clustered in separate enclaves and where their skills were less easily transferable due to local geographic conditions. 

How do people react to a public policy that targets a cultural identity? This paper examines the effects of a top-down secularization-of-education reform on education and religiosity outcomes in Turkey. After the secularization, residents of provinces with higher pre-secularization levels of religiosity were less likely to send their children to secular schools relative to others. I provide evidence consistent with the mechanism that pious parents avoided sending their children to secular schools in order to better transmit their religious identities. They forewent economic benefits of education to achieve this goal and started giving more frequently religious names compared to others after the secularization. Taken together, my findings suggest that education policies that target a cultural identity might result in a backlash and highlight the importance of schooling choices made by parents in transmitting their identity to their children.