How Sensitive are Poverty Measures to the Choice of Equivalence Scale and Unit of Analysis? Evidence from Urban Ethiopia.

Publisert i

Proceedings of the Seventh International Annual Conference on the Ethiopian Economy, Ethiopian Economic Association


The paper studies the effect of the choice of equivalence scale and unit of analysis on poverty measures using data from urban Ethiopia. Four types of equivalence scales and three different units of analysis were considered. The three commonly used FGT indices namely P0, P1 and P2 were computed for the whole sample and for different sub-samples based on geography and socio-economic characteristics of household heads. We found an important effect of the choice of equivalence scale on the poverty measures for the whole sample and for the sub- samples. The magnitude of the effect varies across sub-samples and in some cases rank reversal was observed. This suggests that maximum care should be done while choosing the appropriate equivalence scale. It is also advisable to consider more than one method of equivalence scale to check the robustness of poverty measures to the choice of equivalence scale. Using households as units of analysis consistently understates the level of poverty while the choice between individuals and equivalent individuals does not change the result much. Households should not be used as units of analysis as this could give a misleading result.


By Berhe Mekonnen Beyene
Published June 20, 2011 10:46 AM - Last modified Mar. 6, 2013 12:15 PM