Using survey information about characteristics of personal contacts linked with administrative register information on employment status one year later, the author shows that unemployment survey respondents with many employed acquaintances have a higher job finding rate. Settlement in a socially deprived neighbourhood may, therefore, hamper individual labour market outcomes because of lack of employed contacts. The author investigates this hypothesis by exploiting a unique natural experiment that occurred between 1986 and 1998 when refugee immigrants to Denmark were assigned to municipalities quasi-randomly, which successfully addresses the methodological problem of endogenous neighbourhood selection. Taking account of location sorting, living in a socially deprived neighbourhood does not affect labour market outcomes of refugee men. Furthermore, their labor market outcomes are not affected by the overall employment rate of men living in the neighbourhood, but positively affected by the employment rate of non-Western immigrant men and co-national men living in the neighbourhood. This is strong evidence that immigrants find jobs in part through their employed immigrant and co-ethnic contacts in the neighbourhood of residence and that a high quality of contacts increases the individual's employment chances and annual earnings.