Echelles d'équivalence et différentiel spatial de pauvreté et d'inégalité au Burkina Faso

Based on the household survey carried out by 1994_95, the present study analyzes the relation between the equivalence scales and the spatial poverty in Burkina Faso. Several conclusions are highlighted. Firstly,the estimate of an Engel's curve generates a value of the parameter of scale 2 of 0,53, and a relative child cost for the age group 0_4 years of (=0,6. Secondly, the effect of this new equivalence scale EQ1 – compared to EQ0 where 2 = ( =1– is to raise all measurements of poverty,but the impact appears unequal according to the sectors and regions –reduction in rural poverty compared to urban poverty, and relative variation of the poverty according to geographical areas. Thirdly, the expression of household needs with EQ1 appreciably modifies the characteristics of poverty according to the demography of the groups and areas : (i) inversion of the incidence of poverty of the households according to the sex of the head, female households becoming poorer ;(ii) households having many children are not necessarily poorest –except in the large cities ; (iii) the consideration of EQ1 strongly increases the incidence of poverty in households comprising only the elderly. Fourthly, the statistic 0, testing the null hypothesis of poverty differences according to sectors and areas according to the value of ( and 2, indicates that relative targeting of areas in terms of reduction of poverty with EQ1 is less clear, and diverges somewhat from that which prevails in presence of EQ0. This conclusion is reinforced by the analysis of stochastic second-order dominance. Fifthly, the examination of the sensibility of the incidence of the relative regional poverty according to the variation of the parameters ( and 2 shows : (i)an increase in rural poverty compared to urban poverty almost twice higher when the cœfficient of scale 2 varies from 0,2 to 1 ; (ii) a weak variation of rural-urban relative poverty with (, for a given value of 2 ; (iii) a relative incidence of poverty in terms of national deprivation non-independent of 2. Sixthly, by using the decomposition of the index of Gini proposed by Yitzhaki and Lerman, the study shows that the increase of 2 induces at the same time an increase in the national inequality - even if there is a minimal value of 2 beyond which the variation of Gini changes sign –, and a rise in the weight of between-group inequality – between rural and urban areas – in the explanation of Gini. Furthermore, these trends also characterize each sector - although, for each area, the quasi-totality of Gini is explained by within-group inequality. (Full text in French)