A general method to estimate income distribution functions using information provided in grouped data is presented in this paper. The aggregation may be symmetric (the number of individuals being the same in each interval) or asymmetric, (the number of individuals being different in each interval). This technique enables us to build a worldwide (or national) density function on the basis of national (regional) data which is much more accurate than standard approaches in this type of literature. Our general method is developed in two stages: in the first stage we obtain a Lorenz curve, in the second, using the results of the previous stage an income density function is derived through non-parametric techniques. In addition, several Monte Carlo experiments that prove the good sample properties of our estimator are shown. Finally, two empirical applications of our method of estimation are proposed for both symmetrical and asymmetrical groups.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados