Abstract
The monetary approach is not sufficient to understand the multiple dimension of poverty. The purpose of this research is therefore to measure and analyze multidimensional poverty reduced to a single non-monetary dimension and according to the characteristics of household heads. Based on the non-monetary variables provided by the most recent country survey (QUIBB 2006), we use Multidimensional Correspondence Analysis techniques to construct a Composite Poverty Indicator. The results of the non-poverty index suggest that the poorest are large families and households in rural areas. The deprivation is also more serious in households whose heads are male, aged between 51 and 99 and less educated. The findings are the same for the monetary approach at the poverty line, leading to the conclusion that both types of poverty are quite strongly and positively correlated. Finally, we propose some recommendations for socioeconomic policies.
Acknowledgments
The author thanks the copy editor for her copy editing and three anonymous referees for helpful observations that have been instrumental in improving this paper.
Notes
1 See in brackets the modalities of these three variables: access to toilet (modern toilet, non-modern toilet, no access to toilet), drinking water supply (water tanker truck, untreated water, yard tap, water tap in the home, other water sources), the housing occupancy status (family home, homeowner, tenant).
2 Professional school is a higher education level (i.e. post secondary school level).
3 The first-order dominance stochastic test proves that the results of monetary poverty are generalizable for any monetary poverty threshold (Appendix F).
4 They often consult traditional healers who do not have the skills to diagnose serious illness.