Tiebout sorting, aggregation and the estimation of peer group effects

Growing up in a higher socioeconomic status neighborhood and attending scho ol with socioeconomically advantaged classmates is associated with better a cademic, social, and labor market outcomes. The extent to which this associ ation reflects a causal relationship is much debated, and the conclusions r eached often depend upon the estimation method used to account for the endo geneity of school and neighborhood choice. This paper uses a sample of stud ents from the High School and Beyond Longitudinal Survey to examine whether the use of aggregate county and metropolitan area level data as instrument s for school peer group background ameliorates the problem of endogeneity b ias. The pattern of estimates does not support the belief that aggregation reduces specification error in the estimation of peer group effects. (C) 20 01 Elsevier Science Ltd. All rights reserved.