Violence as a dimension of multidimensional poverty?

Oldiges, Christian; Alkire, Sabina (2018). 'Violence as a Dimension of Multidimensional Poverty?' Paper presented at the annual conference of the HDCA, Buenos Aires, Argentina 2018.


In the Development Economics and Development Studies literature as well as in the wider discourse on what constitutes poverty, it is largely acknowledged that violence (personal security) is a central element of poverty, as well a possible cause and/or consequence. How should this affect poverty measurement? Questions on violence in its various forms such as conflict, crime, and domestic violence are often not included in standardized multi-topic household surveys, nor in surveys on income and consumption. Thus most measures of multidimensional poverty that reflect the joint distributions of deprivations at this time treat violence as a missing dimension.  For example, the internationally comparable Multidimensional Poverty Index covering 10 indicators in health, education, and living standards, violence is missing. In this paper, we marshal existing data on conflict violence [and domestic violence?] to explore this gap, while recognising the potential bias from omitting data on crime. Using GIS coordinates, we merge household-level information on multidimensional poverty with events of political violence collected in the ACLED dataset, then compute an MPI with anew dimension for conflict-related violence (MPI+V) that includes three types of political violence and accounts for the distance to events of violence. With a poverty cut-off of 25 percent, the MPI+V provides information on whether a household is multidimensionally poor in at least one of the four equally weighted dimensions, and if so in how many. Using a conservative approach, which entails an exposure radius of up to 10km, we undertake rank comparisons between the traditional MPI and MPI+V across 28 countries and 300 regions in Africa. Preliminary estimations indicate significantly larger headcount ratios may emerge when violence is included, especially for urban areas. Further, we find that violence against civilians may contribute to overall poverty in countries that do not experience outright wars.Since the mere distance to violence does hide the varying households’ degrees of exposure to violence, we argue for the inclusion of violence modules in household questionnaires.

scroll to top