Measuring Multidimensional Poverty among Children

With the adoption of the Sustainable Development Goals (SDGs) in 2015, the international community affirmed the importance of eradicating child poverty, identifying within SDG Target 1.2 the need to reduce by half the proportion of women, men, and children living in multidimensional poverty. This objective is particularly relevant given that children have been shown to suffer from poverty at disproportionate levels and face the consequences of deprivations during childhood for the rest of their lives. At a minimum, Target 1.2 requires making age decompositions of global poverty statistics the norm. Where resources permit, it is possible to take another step and develop child-specific multidimensional poverty measures that capture the unique attributes of child poverty.

The Alkire-Foster (AF) method can be used in different ways to measure child poverty. First, the subgroup decomposition property of AF measures allows the disaggregation of these measures by age groups. This means that we can estimate the level and intensity of poverty among children (generally defined as individuals younger than 18 years of age) based on AF measures designed to capture poverty at the country level (e.g. the global MPI and national MPIs). As countries’ measures tend to define poverty at the household level, these estimates of child poverty assume a child is poor if she lives in a poor household. Since 2017, OPHI’s updates of the global MPI have included a table with the poverty estimates by age groups. Results have shown that multidimensional poverty is higher among children and that they represent nearly half of all people identified as MPI poor. In particular, children aged 5 through 9 years face significantly higher levels and intensity of multidimensional poverty. Several countries with national MPIs are also starting to disaggregate their poverty figures by age cohorts (e.g. Afghanistan, Bhutan, Chile, El Salvador, and Nepal). Consistent with the global MPI findings, these disaggregations have shown that poverty tends to be higher among children than adults.

Another possible approach to measuring child poverty is to build a standalone child-specific MPI (C-MPI), which identifies poverty at the level of the individual child and considers deprivations that are particularly relevant for children. Such measures can identify poor children living in non-poor households and can reveal deprivations that strike siblings of different ages or genders within the same household differently. There are a few examples of this approach (e.g. Bhutan’s Child MPI and Panama’s MPI for Boys, Girls and Adolescents).  

Although a child-specific measure allows a more detailed focus on the particular situation of children than the disaggregation of the country’s poverty measure, having two distinct measures with different indicators creates challenges in communication and policy applications. To minimise these problems, OPHI researchers are proposing that countries that want to build a C-MPI should ‘link’ it to their national MPI. This ‘linked’ C-MPI, defined at the child level, would include the exact same dimensions and indicators as the national MPI, plus an additional child-specific dimension with age-appropriate indicators that track each child’s individual deprivations throughout her childhood.