Aplicación de pruebas no paramétricas para k-muestras para identificar cambios en la distribución del Ingreso en Ecuador en el periodo 2006-2016 por provincia y actividad económica

Abstract

In this work a description of income distribution is made using kernel density estimation. The period 2006 – 2016 is analyzed using cross-sectional samples from the population not only at the national level but also for the provinces and main economic activities. Non parametric k sample tests based on the kernel density estimation and empirical distribution function area applied to assess whether the observed changes in the distribution are statistically significant. The empirical results suggest that income distribution have changed over time, accompanied with a decreasing in inequality and a subsequent rise the last 2 years, many of the changes are in the localization of the distribution where the EDF tests work well. At the subnational level some changes in shape are identified, this is the case of the province of Chimborazo, where the KDE tests, like norm L1 or AC statistic, reject the null hypothesis.