Suchergebnisse

This paper extends the discussion of the spatial Gini index proposed by Dawkins. Two spatial Gini indices, a nearest neighbour index and a monocentric index, are calculated for a sample of 237 US metropolitan statistical areas (MSAs). For this sample of MSAs, an exploratory analysis is conducted to determine the relationship between the two spatial Gini indices and other traditional measures of residential segregation. Spatial segregation patterns across US census regions and selected US metropolitan areas are also examined. The analysis suggests that the Gini index provides a useful basis for a multidimensional investigation of residential segregation.

Despite the substantial literature devoted to examining the causes of US Black-White residential segregation by race, there is little evidence on the persistence of residential segregation outcomes across generations. This paper examines the following two questions: Do households reside in neighbourhoods with racial compositions that resemble the household head's childhood neighbourhood? Do residential segregation outcomes persist across generations, controlling for household-level determinants of residential location choice? The empirical work relies on household-level data from the US Panel Study of Income Dynamics (PSID) matched to the racial composition of the household's census tract of residence to estimate regression models where the dependent variable is a measure of own-race residential segregation in 1980 and 1990 for each household head in the sample. The independent variables include various contemporaneous household characteristics, characteristics of each household head's parents in 1968 and the lagged 1968 neighbourhood racial composition of the household head's childhood residence. Results from several regression models suggest that households choose to reside in neighbourhoods with racial compositions that resemble the household head's childhood neighbourhood. Observable parental characteristics, particularly measures of parental interracial contact, explain much of the intergenerational persistence in segregation among African American households. Among Whites, the intergenerational persistence of residential segregation remains even in models with an extensive set of controls. Across both racial groups, the neighbourhood social ties of parents living in more integrated neighbourhoods are shown to reduce the propensity for children to choose segregated neighbourhoods upon reaching adulthood. These findings suggest that interracial contact may reduce the persistence of residential segregation over time.

This paper discusses a new measure of residential segregation by race that incorporates spatial proximity among neighbourhoods into the calculation of the index. The basis for the measure is the Gini index of segregation. Unlike other similar measures discussed in recent years, this measure satisfies the 'principle of transfers', is flexible enough to quantify a range of pre-specified spatial patterns of segregation and is easy to compute using spatial statistics software packages. The properties of the index are illustrated using several simple simulations and a case study of non-White-White segregation in Atlanta, Georgia. The application of the index in Atlanta suggests that spatial proximity among adjacent neighbourhoods has a large impact on overall levels of racial segregation.