Income Inequality Comparison through Parametric Modeling
Loading...
Date
2017
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Management and Technology Lahore
Abstract
This study takes into account the determination of income inequality by non-parametric and parametric approach and comparison of the two techniques in the context of Multiple Indicator Cluster survey (MICS) data of Punjab for the years 2007-08 and 2013-14.
Different probability distributions namely Pareto, exponential, generalized Pareto, gamma, Weibull, log-normal, Singh Maddala, Fisk, beta type II and generalized beta type II, have been applied in order to select a suitable distribution for the per capita income of Punjab, Pakistan, for the period of 2007-08 and 2013-14. For this purpose Akaike information criterion (AIC) and Bayesian information criteria (BIC) have been used to choose the best model for the data sets.We find that the Generalized Beta distribution of second kind (GBII) is more appropriate for the given data set. Parameters of the aforementioned distributions have been estimated using the maximum likelihood estimation method. Moreover, Lorenz curves, Gini coefficient (parametric and nonparametric) have been used to measure the income inequality for the data sets. Parametric Gini coefficient for generalized Pareto distribution, not available in literature, has been also derived.
On the basis of parametric Gini coefficient value, it is found that there is a decrease of 0.04 in income inequality in 2013-14 as compared to 2007-08 for the province of Punjab. There also exists a difference in point estimates of Gini coefficient estimated by parametric and non-parametric methods.
Description
Supervised By: Dr. Muhammad Moeen Butt
Keywords
Income Inequality, MS Thesis