Assessing Inequality using Geographic Income Distributions

Ordinarily, an analysis of income differentials based on standard metrics, such as the variance statistic or the gini coefficient, implicitly weights income differentials among different places the same, regardless of whether some pairs of places are more economically interdependent than others. The problem with the assumption that all pairs of places are uniformly interdependent is that changes in those income differentials considered to be less relevant to the inequality concern being addressed may quantitatively obscure acute changes of more relevant differentials.This dissertation has three main chapters. The common aim of each chapter is to incorporate geographic information into a metric's formulation in order to make it more relevant to an explicit concern. Each of the chapters of the dissertation share three objectives: develop a spatial view of inequality based on a concern; incorporate the spatial view into a metric's formulation using a spatial weights matrix; evaluate if the results based on spatial assessments diverge from aspatial ones.An important empirical finding of this research is that a proposed intra-city, inter-race inequality metric registers acute differentials among latino and whiteneighborhoods that an additive decomposition metric does not register. A key conceptual finding is the paradox that spatial inequality metrics formulated for different concerns can register the same change in opposite directions