On the use of the Pareto ArcTan distribution for describing city size in Australia and New Zealand

Abstract

The circular inverse of the tangent function is used to simply derive a generalization of the Pareto distribution, the Pareto ArcTan (PAT) distribution. This model includes as limiting cases Pareto and Zipf distributions. This new probabilistic family is used to describe city size data for Australia and New Zealand. Urban agglomerations of these two countries presents similar features, a few large metropolitan areas with a steadily increasing population in the last years and many small cities. The PAT distribution improves the performance of other traditionally used models in urban agglomeration economics such as the classical Pareto, lognormal and the recently proposed Pareto positive stable.