Fig. 5: County-level correlations between upward income mobility and neighbourhood characteristics.

a, County-level univariate correlations of upward income mobility with economic connectedness (EC) and other county characteristics obtained from external datasets (see Supplementary Information A.5 for details). Upward income mobility is obtained from the Opportunity Atlas⁷² and is measured as the predicted household (HH) income rank in adulthood for children in the 1978–1983 birth cohorts with parents at the 25th percentile of the national income distribution. Income segregation is defined using a Theil (entropy) index⁸¹. Racial segregation is defined using Theil's H-index across four groups (white, Black, Hispanic, other). See Supplementary Information A.5.1 for details. The Gini coefficient is defined as the raw Gini coefficient estimated using tax data minus the income share of the top 1% to obtain a measure of inequality among the bottom 99% in each county¹⁰. The rest of the variables are all obtained from the Opportunity Atlas⁷². Test scores are measured in third grade, which includes children who are 8 to 9 years old. b, Estimates from a single multivariable regression of upward mobility on a subset of variables from a, with both the outcome and dependent variables standardized to have a mean of zero and a standard deviation of one. The variables used in b are the seven variables from a that have the largest univariate correlations with upward mobility (except the share of households above the poverty line, which is highly correlated with median household incomes), which include all of the strongest predictors of mobility identified in prior work¹⁰. All correlations and regressions are weighted by the number of children in each county whose parents have below-national-median income. Intervals represent 95% confidence intervals calculated using standard errors clustered by commuting zone.