Over the last decade, poverty mapping methodologies have been developed in- dependently by econometricians and statisticians working in small area estimation. Small area estimation of key economic variables, including poverty, has been at the centre of methodological and applied work in official statistics. This is reflected by the fact that the European Commission, via the 7th Framework, has recently funded two research projects (SAMPLE and AMELI) that aimed at developing novel methodology for estimating small area poverty indicators. Only very recently, however, there have been attempts to understand the differences between some of the alternative poverty mapping methods. This paper sets to discuss the properties of three poverty mapping method- ologies namely, (a) the World Bank approach-ELL (Elbers et al, 2003), (b) the Empirical Best Predictor approach-EBP (Molina and Rao, 2010) and (c) the M- quantile approach (Chambers and Tzavidis, 2006; Marchetti et al, 2010) when used for estimating key poverty indicators such as the Head Count Ratio and the Poverty Gap (Foster et al, 1984). In particular, we focus both on properties of point and MSE estimators, which under all three methods are respectively derived by means of Monte-Carlo simulation and bootstrap. The comparisons are carried out by using a series of model-based simulations as well as real applications. Model-based simulations are designed under different scenarios for the distribution of the model error terms and the area-specific sample sizes. Finally, the poverty mapping methodologies are applied to two real datasets. In the first application data come from the European Survey of Income and Living Conditions in Italy and the aim is to estimate the incidence of poverty and the poverty gap in Italian provinces in Tuscany, Lombardia and Campania. In the second application data come from the ENIGH survey in Mexico and the aim is to estimate the incidence of poverty and the poverty gap for municipalities in the State of Veracruz.
A Comparison of the ELL, EBP and M-Quantile Poverty Mapping Methodologies
MARCHETTI, STEFANO;GIUSTI, CATERINA;PRATESI, MONICA
2011-01-01
Abstract
Over the last decade, poverty mapping methodologies have been developed in- dependently by econometricians and statisticians working in small area estimation. Small area estimation of key economic variables, including poverty, has been at the centre of methodological and applied work in official statistics. This is reflected by the fact that the European Commission, via the 7th Framework, has recently funded two research projects (SAMPLE and AMELI) that aimed at developing novel methodology for estimating small area poverty indicators. Only very recently, however, there have been attempts to understand the differences between some of the alternative poverty mapping methods. This paper sets to discuss the properties of three poverty mapping method- ologies namely, (a) the World Bank approach-ELL (Elbers et al, 2003), (b) the Empirical Best Predictor approach-EBP (Molina and Rao, 2010) and (c) the M- quantile approach (Chambers and Tzavidis, 2006; Marchetti et al, 2010) when used for estimating key poverty indicators such as the Head Count Ratio and the Poverty Gap (Foster et al, 1984). In particular, we focus both on properties of point and MSE estimators, which under all three methods are respectively derived by means of Monte-Carlo simulation and bootstrap. The comparisons are carried out by using a series of model-based simulations as well as real applications. Model-based simulations are designed under different scenarios for the distribution of the model error terms and the area-specific sample sizes. Finally, the poverty mapping methodologies are applied to two real datasets. In the first application data come from the European Survey of Income and Living Conditions in Italy and the aim is to estimate the incidence of poverty and the poverty gap in Italian provinces in Tuscany, Lombardia and Campania. In the second application data come from the ENIGH survey in Mexico and the aim is to estimate the incidence of poverty and the poverty gap for municipalities in the State of Veracruz.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
