Comparing global trends in multidimensional and income poverty and assessing horizontal inequalities

The 2030 Agenda has provided new impetus to two facets of the struggle for poverty alleviation, which is a central goal of the international development community. First, poverty is no longer viewed strictly in monetary terms, but rather as a multidimensional phenomenon. Second, the need to reduce poverty for different social groups and not just at the aggregate, national level is explicitly recognised. Against this background, this paper has three objectives: (1) to analyse the trends in multidimensional poverty in low- and middle-income countries, (2) to explore rural-urban differences in poverty over time, and (3) to assess the validity of the claim that there has been a feminisation of poverty. The analysis relies on a new indicator of multidimensional poverty, the Global Correlation Sensitive Poverty Index (G-CSPI), that incorporates three key components: education, employment and health. The G-CSPI has several methodological advantages over existing measures, including that it is an individual rather than a household-level measure of poverty, which is crucial for gender-disaggregated analysis. Regarding aggregate trends, this paper shows that both income poverty and multidimensional poverty fell between 2000 and 2012. However, the decline in (extreme) income poverty in percentage terms was twice as large as the decline in multidimensional poverty. There is significant heterogeneity in the results across regions. Multidimensional poverty declined the most in Asia, converging towards the relatively low levels of Latin America and Europe, while sub-Saharan Africa’s slow progress further distanced it from other regions. These findings point to the existence of poverty traps and indicate that more efforts are needed to eradicate poverty. Regarding the urban-rural comparison, our analysis shows that poverty is predominantly a rural phenomenon: the rural G-CSPI was more than four times the urban G-CSPI. This difference remained nearly constant over time. As for the third objective, we find no gender bias in 2000 at the global level. This contrasts with the claim made in 1995 in Beijing that 70 per cent of the poor were women. However, we find that multidimensional poverty declined more among men (-18.5 per cent from 2000) than women (-15 per cent), indicating a process of feminisation of poverty. This was triggered by the decline in employment poverty, which was much slower among women. As most existing studies conclude that there was no evidence of the feminisation of poverty, this finding is new to the literature.