We investigate the relationship between neighborhood income composition and income trajectories of adults, employing annual panel data from Stockholm over the 1991–2008 period and multiple measures of neighborhood income mix. We advance the human geography literature in three ways by quantifying neighborhood effects that: (1) are unusually precise due to our large sample size; (2) are arguably causal and unbiased due to the econometric techniques employed; (3) are potentially heterogeneous, varying according to gender, income group, and ethnicity. Our innovative, fixed‐effect change modeling indicates that neighborhood income mix affects subsequent one‐ and five‐year income trajectories of residents in highly heterogeneous ways according to gender, income and ethnicity, and for some groups this effect is substantial. The evidence supports on Pareto improvement grounds a social mix policy that attempts to reduce the incidence of lower‐income dominant neighborhood environments and replace them with more mixed or middle‐income dominant ones.