Effects of neighborhood labeling on student performance and sorting

We study how labeling of neighborhoods affects the performance and sorting of young residents at the junction of enrolling in high school. Exploiting the release of a much publicized official situation report on carefully delineated troubled neighborhoods in Sweden we estimate the effect of this unanticipated and well-documented negative information shock of neighborhood quality. Our research strategy compares the change in outcomes of students living in neighborhoods that were listed in the report with those residing in non-listed ones. To shed light on mechanisms, we focus on one hand on self reported perceptions and beliefs from national school survey data while on the other hand empirically ruling out any potential simultaneous place-based investments by using geocoded longitudinal employer-employee data covering the entire Swedish population. We show that the negative information shock: (i) does not affect compulsory school performance of students living in a neighborhood listed as "troubled" in the report, (ii) increases sorting of students living in these troubled neighborhoods into lower quality high schools located further away from their homes, (iii) has a greater effect on students from relatively higher socioeconomic background, (iv) affects treated students' perception of fairness and the extent to which effort pays off. Our results are consistent with models of neighborhood labeling, wherein residents in neighborhoods with a negative public image adjust their behavior in response to social image concerns.