Egohoods that segregate : immigrants, social contexts, and income penalties

This paper looks at the causal impact of residential segregation on income among migrants in Brescia, a wealthy industrial city of about 200,000 inhabitants in North‐Eastern Italy, where almost one citizen out of five is born abroad. Using a new register‐like dataset including economic and demographic microdata from administrative sources, we geo‐localised each migrant household and created for each individual an “egohood” by imposing buffer circles with the same diameter (250 or 500 m) around his or her residential location. Within this spatial area of reference, we calculated indices of ethnic segregation and exposure to Italian natives, thus measuring the individual network of potential contacts with co‐ethnics and with Italians. Following social networks and social capital theories, we expected ethnic segregation to have a negative impact and exposure to natives to have a positive impact on income and tested our hypotheses with a set of ordinary least squares and two‐stage least squares with instrumental variables regression models. Results showed a robust and negative effect of ethnic segregation on income, whereas a small and positive effect of exposure to natives was not robust to endogeneity checks. We also found substantial heterogeneity among migrant groups defined by macroareas of origin: Exposure to Italians was found to have the expected impact only among migrants coming from Eastern Europe and from the Middle East and North Africa.