Aplicación de pruebas no paramétricas para k-muestras para identificar cambios en la distribución del ingreso en Ecuador 2006-2016.

Abstract

This paper analyzes the personal income distribution at different levels: national, provincial and by economic activity for the 2006-2016 period, based on information from household surveys. To test whether changes in the distribution of personal income over time are statistically significant, non-parametric tests for k samples, based on the Empirical Distribution Function (EDF) and the Kernel Density Function (KDE) are applied. The results show that, indeed in the time period analyzed, there were changes in the distribution of personal income, mainly in location, rather than in form, where the tests based on the EDF perform well. Meanwhile, at the territorial level, changes in the shape of the distribution were evidenced, as in the case of Chimborazo, where the tests based on the KDE, such as the L1 norm or the Common Area, rejected the null hypothesis.