Abstract:

Recent research documents a causal impact of place on the long-run outcomes of children.  However, little is known about which neighborhood characteristics are most important, and at what scale neighborhood effects operate.  By using the random assignment of public housing along with administrative data from Denmark, we get inside the "black box" of neighborhood effects by defining neighborhoods using various characteristics and scales.  Results indicate effects on mental health and especially education are large but local, while effects on drug possession operate on a much broader scale.  Additionally, unemployment and education are better predictors of outcomes than neighborhood income.  

Stephen B. Billings (University of Colorado  Boulder)  Mark Hoekstra (Texas A&M University; NBER)  Gabriel Pons Rotger (VIVE)