Mansouri, Brahim, and Fouzia Ejjanoui (2009). "Multidimensional Measurement of Poverty in Morocco: The Case of Urban Poverty in the City of Marrakech" Paper presented at the 6th annual conference of the HDCA, 10-12 September 2009, Lima, Peru.

Recent studies of poverty in Morocco have heavily relied on the monetary approach adopted by the World Bank. To characterize and measure the poverty phenomenon in Morocco, the Direction de la Statistique in Rabat uses the households’ expenditures as the leading variable. Well, this approach has proved its shortcomings in Morocco and the whole developing world as well, especially because of its particular mono-dimensionality. Indeed, a universal consensus has been occurred about the fact that poverty is somewhat a multiple privation. Thus, it would be interesting to take into consideration all monetary and non-monetary dimensions when measuring and analyzing poverty. The present paper project seeks to measure multidimensional poverty in the Moroccan city of Marrakech, using raw data from the 2000-2001 National Survey on Households’ Expenditures (Enquête Nationale sur les Dépenses des Ménages), and the 2004 General Census of People and Housing (Recensement Général de la Population et de l’Habitat). More precisely, we aspire to constructing a composite indicator of poverty for each household in the city. Two mean objectives have to be achieved through this research: i) analyzing the interaction between monetary and non-monetary poverty; and ii) to elaborate a multidimensional poverty map in Marrakech, using our multidimensional approaches. To do this, we have used two advanced empirical methodologies: the Multiple Components Analysis (Analyse des Correspondances Multiples), and the Classification Ascendante Hiérarchique. Our empirical results reveal that the monetary and non-monetary measures of poverty are weakly correlated. Results show also that certain households who are not subjected to monetary poverty suffer from multidimensional poverty. As far as the geographical distribution of monetary poverty through the various city districts is concerned, we have observed a strong disparity. The gap in terms of poverty between the richer district of Gueliz and the poorer district of Annakhil is estimated to be around 18.4 percent. Our empirical results from the multidimensional poverty map are largely different from those obtained by the monetary poverty map elaborated by official studies, especially those of the Haut Commissariat au Plan (Rabat, Morocco). On the basis of our empirical results, we confirm the shortcomings of the monetary approach to poverty in identifying the poor and its tendency to under-evaluate the phenomenon. When they rely heavily on the monetary approach, poverty alleviation policies would very probably be inefficient. Even when we assume that such policies are well-targeted, they would likely not benefit for the households considered as non-poor according to the monetary approach, while they are seen to be as poor following the multidimensional approach. This permits to conclude that it is necessary to take into account all dimensions of poverty, not only the monetary dimension, based on households’ expenditures (or income). As we have done, all available surveys and censuses should be exploited in order to grasp all the dimensions of poverty. The marginal cost of multidimensional approaches to poverty is low while their informational gains are particularly considerable.