Abstract
This note examines the level of polarization in the US state per capita income distribution during the period 1929 to 2005, using, for the first time in this context, the set of measures proposed by Esteban and Ray (Citation1994) and Esteban et al. (Citation2007). Our results show that income polarization has decreased over the sample period, regardless of the number of groups considered in the study or the value assigned to the polarization sensitivity parameter. In addition, the analysis carried out reveals the existence of a statistically significant relationship between the evolution of spatial autocorrelation and income polarization, which highlights the relevance of physical-geography spillover effects in explaining the observed changes in the distribution under consideration between 1929 and 2005.
Notes
1 This choice is due to the fact that, as mentioned above, the formulation of PER is similar to that of the Gini index. In turn, the second term in expression (3) is the difference between two Gini indices. In empirical analyses, therefore, it is reasonable to set β at a value equal to the unit (Duro, Citation2005).
2 Additional information on the polarization of the simplified representations of the original distribution, and the degree of within-group dispersion is available upon request.
3 Inference is based on the permutation approach with 10 000 permutations.