Table 2.
Associations between upward income mobility, EC and other neighbourhood characteristics
| EC versus median income and poverty rates | ||||||||
|---|---|---|---|---|---|---|---|---|
| Dependent variable | Upward income mobility | |||||||
| Counties | ZIP codes | |||||||
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) | |
| Median income | 0.345*** | –0.006 | – | – | 0.574*** | 0.209*** | – | – |
| (0.045) | (0.069) | – | – | (0.023) | (0.029) | – | – | |
| Poverty rate | – | – | -0.299*** | 0.142** | – | – | –0.543*** | -0.195*** |
| – | – | (0.070) | (0.069) | – | – | (0.052) | (0.054) | |
| Economic connectedness | – | 0.649*** | – | 0.732*** | – | 0.548*** | – | 0.568*** |
| – | (0.058) | – | (0.043) | – | (0.038) | – | (0.033) | |
| Observations | 2,984 | 2,984 | 2,984 | 2,984 | 24,165 | 24,165 | 24,165 | 24,165 |
| R2 | 0.119 | 0.418 | 0.089 | 0.430 | 0.330 | 0.496 | 0.295 | 0.496 |
| EC versus segregation and inequality | ||||||
|---|---|---|---|---|---|---|
| Dependent variable | Upward income mobility | |||||
| (1) | (2) | (3) | (4) | (5) | (6) | |
| Income segregation | –0.173*** | -0.071 | – | – | – | – |
| (0.053) | (0.054) | – | – | – | – | |
| Racial segregation | – | – | –0.212** | -0.027 | – | – |
| – | – | (0.088) | (0.086) | – | – | |
| Income inequality (Gini coefficient) | – | – | – | – | –0.449*** | -0.103 |
| – | – | – | – | (0.084) | (0.091) | |
| Economic connectedness | – | 0.601*** | – | 0.604*** | – | 0.577*** |
| – | (0.044) | – | (0.054) | – | (0.063) | |
| Observations | 1,820 | 1,820 | 1,821 | 1,821 | 2,741 | 2,741 |
| R2 | 0.034 | 0.413 | 0.051 | 0.408 | 0.207 | 0.424 |
| EC versus share of Black residents | ||||||||
|---|---|---|---|---|---|---|---|---|
| Dependent variable | Upward income mobility for Black individuals | Upward income mobility for white individuals | ||||||
| Counties | ZIP codes | Counties | ZIP codes | |||||
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) | |
| Share of Black individuals | –0.158** | 0.078 | –0.204*** | –0.014 | –0.128** | 0.151** | –0.250*** | 0.035* |
| (0.068) | (0.076) | (0.057) | (0.071) | (0.057) | (0.067) | (0.018) | (0.018) | |
| Economic connectedness | – | 0.502*** | – | 0.468*** | – | 0.582*** | – | 0.631*** |
| – | (0.095) | – | (0.083) | – | (0.051) | – | (0.027) | |
| Observations | 1,885 | 1,885 | 11,147 | 11,147 | 2,982 | 2,982 | 24,020 | 24,020 |
| R2 | 0.025 | 0.222 | 0.042 | 0.224 | 0.016 | 0.277 | 0.063 | 0.380 |
This table presents estimates from ordinary least squares (OLS) regressions of upward income mobility on economic connectedness (EC) and other area-level characteristics. Upward income mobility is obtained from the Opportunity Atlas72 and is measured as the predicted household income rank in adulthood for children in the 1978–1983 birth cohorts with parents at the 25th percentile of the national income distribution. EC is twice the share of above-median-SES friends among below-median-SES people. We standardize every dependent and independent variable to have a mean of zero and variance of one (weighted by the number of children with below-median parental income in the county). For ‘EC versus median income and poverty rates’ and ‘EC versus segregation and inequality’, the dependent variables are upward mobility pooling all racial and ethnic groups72, and regressions are weighted by the number of children with below-median parental income. ‘EC versus median income and poverty rates’ presents regressions at both the county and ZIP code levels, with median household income and poverty rates by county and ZIP code obtained from the 2000 Census. In ‘EC versus segregation and inequality’, all regressions are estimated at the county level. 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. Gini coefficients are 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. ‘EC versus share of Black residents’ presents regressions at both the county and ZIP code levels. The dependent variables are upward mobility estimates for Black and white individuals separately74. ‘Share of Black individuals’ is from the 2000 Census. All regressions in this section are weighted by the race-specific number of children with below-median parental income in the county. See Supplementary Information A.5 for further details on data sources for neighbourhood-level characteristics. Standard errors (reported in parentheses) are clustered at the commuting zone level. Asterisks indicate the level of significance: *10%, **5%, ***1%.