Measuring U.S. Spatial Poverty Concentration: Methodological Implications for Public Polic

This study focused on the spatial measurement of urban neighborhood poverty concentration in the largest 100 metropolitan regions between 1980 and 2010. First, we used global models of OLS, spatial autocorrelation, and spatial regression methods to model and analyzed neighborhood poverty across different racial/ethnic groups. We provided broad summary statistics on urban neighborhood poverty across these groups; and we analyzed and measured patterns and trends of urban poverty rates from 1980 to 2010.;Then using GWR model, we analyzed spatial variation in neighborhood poverty concentration for the year 2010. Using the sample of the four selected metropolitan regions (Detroit, Dallas, Baltimore, and Chicago) first, the OLS global model was fitted to estimate the relationships between neighborhood poverty concentration and the independent variables in these regions. Then the GWR model was then calibrated to investigate and capture spatial non-stationarity in the local parameter estimates of each one of the four selected metropolitan regions. Urban poverty concentration is measured, mapped, and high/low poverty neighborhoods were identified. We recommended the implementation of a hybrid system of poverty mitigation policy which should include both race and place components.