Park, SeonyoungShin, Donggyun2019-03-202022-07-122019-03-202022-07-12202020202230-2603https://ir.wgtn.ac.nz/handle/123456789/20926Analysis of US county per capita incomes from 1970 to 2017 reveals the emergence of bipolarizing distribution dynamics from the early 1990s. This bipolarization process is characterized by the vanishing middle income counties mostly joining the high end of the distribution of county incomes. Cross-county differences in education and industry composition contribute to the bipolarization, but government transfers effectively reverse it. The results for these recent decades weakly support the two-club convergence hypothesis. A simulation of various nonlinear income growth dynamics and corresponding distributional dynamics reveals certain conditions on growth patterns for income bipolarization.pdfen-NZhttp://www.victoria.ac.nz/sef/research/sef-working-papersbipolarizationgrowth dynamicsdistributional dynamicsindexnon-parametric estimationText