Multidimensional Poverty

In addition to an indicator-by-indicator dashboard analysis, this study also estimates a multidimensional measure of poverty to investigate the experience of simultaneous deprivations following Alkire and Foster (2011). In brief, this method counts deprivations for a set of dimensions and indicators.

An individual is considered to experience multidimensional poverty if the number of deprivations of the individual exceeds a set threshold. Details on the calculation of this measure are included below. H is the multidimensional poverty headcount and gives the percentage of the population who experiences multidimensional poverty or multiple deprivations. Dimensions are weighted and wj is the weight of dimension j. There are different possible methods for setting up weights, for instance, asking people’s opinions or using the observed distribution of successes or deprivations (Decancq and Lugo 2013).

In this report, as is often done in multi-dimensional poverty research, all dimensions were considered equally important and were given equal weights (each has a weight of 1) and when more than one indicator was used within a dimension, indicators were equally weighted within the dimension. For instance, for the health dimension with two indicators, each indicator weighs ½.

According to the method laid out in Alkire and Foster (2011), each individual i has a weighted count of dimensions where that person achieves deprivations (c_i) across all measured dimensions: 0≤ c_i ≤ d where d is the number of dimensions; with equal to one if individual i has a deprivation in dimension j, and zero otherwise. Let q_i be a binary variable equal to one if the person is identified as deprived, and to zero otherwise. A person is identified as experiencing multidimensional poverty if the person’s count of deprivations is greater than some specified cutoff (k):

if c_i > k, then q_i = 1; if c_i ≤ k, then q_i = 0

In this study, k=1.

The multidimensional poverty headcount or the share of adults experiencing multidimensional poverty H is then the number of persons in multidimensional poverty ( q_i ) divided by the total population (n): H=q/n

Dimensions and indicators are laid out in Table 1 below.

Based on the information available in the datasets under study, four dimensions and eight indicators were selected for the calculation of the multidimensional poverty measure. The four dimensions are: education, work, health, and standard of living. Each has a weight of 1 and when more than one indicator was used within a dimension, indicators were equally weighted within the dimension.

Education is measured through an indicator of educational attainment for adults. Work is measured through work status. Health is measured with two indicators and each has a weight of ½: access to safely managed drinking and sanitation services. Standard of living is measured through four indicators with each a weight of ¼: clean fuel, electricity, adequate housing and asset ownership.

The cutoffs for the dimensions are as follows: if a person (1) has less than primary education; (2) is not working; (3) lives in a household without safely managed drinking water; (4) lives in a household without safely managed sanitation services; (5) lives in a household without clean cooking fuel; (6) lives in a household without adequate housing, i.e. without adequate walls, floor and roof; (7) lives in a household without assets.

More details on how indicators are defined is in Method Brief 2. For DHS datasets, we did not include the work indicator in the multidimensional poverty analysis due to small sample sizes at the subnational level.

References

Alkire, S. and Foster, J. (2011). Counting and multidimensional poverty measurement. Journal of Public Economics 95(7–8):476–87.

Decancq, K. and Lugo, M. A. (2013). Weights in multidimensional indices of wellbeing: an overview.Econometric Reviews, 32, 7-34.