A Multi-Country Analysis of Multidimensional Poverty in Contexts of Forced Displacement

Yeshwas Admasu, Sabina Alkire, Uche Eseosa Ekhator Mobayode, Fanni Kovesdi, Julieth Santamaria, Sophie Scharlin-Pettee


This paper develops a Multidimensional Poverty Index (MPI) to examine patterns of multidimensional poverty among IDPs and refugees, with comparisons to host populations, in five African countries. The MPI is disaggregated to analyze variations in deprivation by displacement status of the household and gender of the household head. The analysis draws on household survey data from Ethiopia, Nigeria, Somalia, South Sudan, and Sudan[1].

The MPI measure used in this paper includes 15 indicators covering four non-monetary dimensions of poverty: (1) education (years of schooling, school attendance); (2) health (food security, pregnancy care, physical safety, early marriage); (3) living standards (garbage disposal, drinking water, electricity, cooking fuel, housing, sanitation); and (4) financial security (unemployment, legal identification, bank account). For each indicator, households are assessed as deprived or non-deprived according to specific criteria. The calculation of the MPI uses equal weights for all four dimensions and, within each dimension, all indicators have equal weights. Households deprived in at least half of the weighted indicators are identified as multidimensionally poor. The authors calculate ‘incidence’ or headcount ratio, which is the proportion of the population who are multidimensionally poor; ‘intensity’, which is the average share of weighted deprivations experienced by the poor; and an adjusted headcount ratio, the MPI, which is a product of the incidence and intensity.

Main findings:

  • The incidence of multidimensional poverty varies across countries. Overall incidence is highest in Somalia, where 63 percent of IDPs and 45 percent of the host population are identified as multidimensionally poor, while the lowest overall incidence is found in northeastern Nigeria, where 23 percent of IDPs and 17 percent of the host population are multidimensionally poor.
  • In all countries, displaced populations have higher incidence of multidimensional poverty, with the largest gap in countries hosting refugees or IDPs in camp settings. Differences in the incidence of multidimensional poverty between displaced and host populations are highest in Ethiopia and Sudan, where the surveys were carried out in camps. Refugees in Ethiopia have rates of multidimensional poverty 33 percentage points higher than the host population, while IDPs in Sudan have rates of multidimensional poverty 34 percentage points higher than the host population. The smallest significant differences in incidence between displaced and host populations were found in South Sudan (15 percentage points) and Somalia (19 percentage points).
  • The proportion of people identified as multidimensionally poor and deprived in a given indicator is higher among displaced populations for most indicators—but there is variation across countries in which indicator is the most salient. In Ethiopia, the largest gap between displaced and host populations are for indicators on access to bank accounts and cooking fuel. In Somalia it is years of schooling, in Sudan it is electricity, in South Sudan it is drinking water, and in Nigeria it is legal identification.
  • The percentage contribution of each indicator to multidimensional poverty also varies between displaced and host populations. For example, among refugees in Ethiopia, lack of a bank account is the largest contributor to multidimensional poverty, while among host communities, the largest contributor is years of schooling. Physical safety, early marriage, and legal identification also contribute more to multidimensional poverty among refugees in Ethiopia than among host communities.
  • Multidimensional poverty varies depending on the gender of the household head. Those in female-headed households have a much higher MPI compared to those in male-headed households in Ethiopia, South Sudan and Sudan, while the opposite is true for Somalia, where people living in male-headed households have a higher MPI. In Nigeria, there is no statistically significant difference between the MPIs of the two groups. The gender of the household head matters also in relation to the incidence of multidimensional poverty, i.e., the proportion of the population who are multidimensionally poor: those in female-headed households are 39 percentage points poorer in Ethiopia, and 18 and 10 percentage points poorer in South Sudan and Sudan, respectively.
  • Disaggregating multidimensional poverty outcomes by both gender of the household head and displacement status is revealing. In Ethiopia, the MPI and incidence of multidimensional poverty in the refugee population are much higher for those living in female-headed households, compared to those living in male-headed households, but there are no statistically significant differences by gender of the household head in the host community. In South Sudan, individuals living in female-headed households are more likely to be multidimensionally poor compared to those in male-headed households, both in the IDP and host population. A counterintuitive result is found in Somalia, where the MPI and incidence of multidimensional poverty is higher for people living in male-headed households, in both IDP and host populations. And in Sudan and Nigeria, there are no statistically significant differences in multidimensional poverty outcomes between male- and female-headed households in both IDP and host communities.
  • Drivers of observed gender-based differences in multidimensional poverty vary across countries. In general, in countries where female-headed households are worse off, the gender gaps in access to legal identification, access to a bank account, early marriage, and physical safety are larger than the gap in other indicators. In Somalia, where male-headed households are poorer, the gap in education is a much larger contributor to the gender gap, followed by the gap in access to legal identification.
  • Households that depend mostly on women’s earnings seem more likely to be in multidimensional poverty. For example, in Ethiopia, multidimensional poverty rates among female-single earners and multiple-female earners in refugee households are the highest (52 percent and 57 percent, respectively). These rates surpass even the multidimensional poverty rate of households without earners. In contrast, 23 percent of refugee households whose income depends on a single-male earner and 16 percent of those that depend on multiple-male earners are poor.
  • In all countries, female-headed IDP/refugee households have both a higher incidence of multidimensional poverty and a higher overall MPI than female-headed households that are not displaced. Additionally, for female-headed households in camps (Ethiopia, Sudan) multidimensional poverty is more prevalent and more intense than for female-headed households living outside camps.
  • Overall, monetary poverty and multidimensional poverty measures identify different households as poor. Despite these differences, people living in refugee/IDP households are consistently poorer according to both the monetary and the multidimensional poverty measures. Additionally, households with a single female earner or no earners are consistently identified as poorer according to each of the measures.

The authors argue that the results offer compelling evidence in support of a more disaggregated multidimensional poverty analysis. The findings indicate that differentiation by gender of the household head, earnings profile, and displacement status of the household have important implications for policy and targeting.




[1] In Ethiopia, the Skills Profile Survey (2017) sampled refugees in and around camps in the Tigray, Afar, Gambella, Benishangul Gumuz, and Somali regions. In Nigeria, the IDP Survey (2018) sampled IDPs and host communities in six northeastern states (Adamawa, Bauchi, Borno, Gombe, Taraba, and Yobe). In Somalia, the High Frequency Survey (2017) sampled IDPs and host communities in secure parts of the country. In Sudan, the IDP Profiling Survey (2018) sampled IDPs and neighboring host communities in the Abu Shouk and El Salam camps, in Al-Fashir. And in South Sudan, the High Frequency Survey Wave 4 (2017) sampled IDPs and host communities in urban areas of seven of the ten pre-war states (Western Equatoria, Central Equatoria, Eastern Equatoria, Northern Bahr-El-Ghazl, Western Bahr-El-Ghazal, Warrap, Lakes state).


Pin It on Pinterest