A measurement of spatial disparity: The case of income inequality

Traditionally, inequality measurements have been designed only for non-spatial applications, despite the argument that inequality is manifested in social space (and may be perceived spatially), Presented here is an approach to understanding urban inequality in terms of the spatial distribution of population and income, This is done with a general definition of spatial disparity (implying significant differences between neighbouring parcels); spatial disparity exists between contiguous parcels even when similar values are spatially clustered (as in the case of income), A family of spatial disparity measures is proposed: the basic measure of 'neighbourhood disparity' uses information on contiguity to arrive at individual disparity values for individual parcels, and an indexed value for an area, The process of neighbour identification is made simpler by using topologically structured spatial databases used in vector-based CIS, This methodology is used on 1990 census data to analyse disparities in the spatial distribution of income in Philadelphia to illustrate and compare the use, the basic measure and its weighted versions, and to identify and map general patterns of urban spatial income disparity, It is found that spatial income disparity is manifested in rings around the CBD, and that the distribution of black incomes is very heterogeneous (although the generalisability of both conclusions may be questioned).