Working Paper: NBER ID: w9041
Authors: Kerwin Kofi Charles; Patrick Kline
Abstract: This paper posits that individuals can more easily form social connections with persons of the same race. If true, the greater the incidence among his neighbors of persons of his race, the more likely an individual is to make neighborhood social capital connections, and the more likely he is to engage in activities which require it. The paper tests this idea using an indicator of individual social capital never previously studied: whether the person uses a carpool to get to work. We identify exogenous variation in adult neighborhood racial makeup arising from the racial makeup of the state in which the person was born in the year that he was born, and relate this exogenous portion of adult neighborhood racial composition to individual carpooling propensity using a TSLS approach. The results from this analysis, and from robustness tests which focus on neighborhoods with virtually identical racial distributions, show evidence of strong cross-racial relational difficulties, but interestingly, only for particular pairs of racial groups.
Keywords: social capital; carpooling; racial composition
JEL Codes: Z13
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| neighborhood racial composition (R23) | carpooling behavior (D16) |
| fraction of neighbors of the same race (J15) | social capital accumulation (Z13) |
| social capital accumulation (Z13) | carpooling behavior (D16) |
| relational costs associated with race (J15) | propensity to carpool (R48) |
| cross-racial relational difficulties (J15) | social capital investment (E22) |