ON THE ESTIMATION OF GROWTH AND INEQUALITY ELASTICITIES OF POVERTY WITH GROUPED DATA

After decades of intensive research dedicated to efficient and flexible parametric statistical distributions, the lognormal distribution still enjoys, despite its empirical weaknesses, widespread popularity in the applied literature related to poverty and inequality analysis. In the present study, we emphasize the drawbacks of this choice for the calculation of the elasticities of poverty. For this purpose, we estimate the growth and inequality elasticities of poverty using 1,132 income distributions, and 15 rival assumptions on the shape of the income distributions. Our results confirm that the lognormal distribution is not appropriate in most cases for the analysis of poverty: the magnitude of the elasticities is generally overestimated and the estimation of the relative impact of growth and redistribution on poverty alleviation is biased in favor of the growth objective. Copyright 2009 The Author. Journal compilation International Association for Research in Income and Wealth 2009.