Publikation Poverty Estimation Methods: a Comparison under Box-Cox Type Transformations with Application to Mexican Data

Eradication of poverty is one of the most important Millennium Development Goals defined by the United Nations. One approach to studying this is by using small area (SA) methods. SA models combine survey and census data for producing estimates of poverty indicators at geographical levels where direct estimation is either not possible due to the lack of sample observations or very imprecise. The objective of this work is twofold. First, to analyze how the performance of this method is affected by departures rom normality. Second, to explore how Box-Cox transformations can assist with improving the validity of model assumptions and the precision of SA prediction. This work involves extensive model-based simulations. Finally, the spatial distribution of poverty in Mexico by using real data is analyzed.

Auszug aus der Publikation "WISTA - Wirtschaft und Statistik", 3/2017

Autorin: Natalia Rojas Perilla