Does neighborhood income mix affect earnings of adults? New evidence from Sweden

This paper contributes to the literature on obtaining unbiased estimates of neighborhood effects, explored in the context of a centralized social welfare state. We employ a longitudinal database comprised of all working age adults in metropolitan Sweden 1991-1999 to investigate the degree to which neighborhood income mix relates to subsequent labor incomes of adults and how this relationship varies by gender and employment status. We control for unobserved, time-invariant individual characteristics by estimating a first-difference equation of changes in average incomes between the 1991-1995 and 1996-1999 periods. We further control for unobserved time varying characteristics through an analysis of non-movers. These methods substantially reduce the magnitude of the apparent effect of neighborhood shares of low-, middle- and high-income males. Nevertheless, statistically and substantively significant neighborhood effects persist, though relationships are nonlinear and vary by gender and employment status. Males who are not fully employed appear most sensitive to neighborhood economic mix in all contexts. (C) 2007 Elsevier Inc. All rights reserved.