Applying expert knowledge to measure multidimensional rural poverty in Chittagong (Bangladesh)

  • Authors:
    Melania Salazar-Ordóñez, Lorenzo Estepa Mohedano, Rosa Cordón-Pedregosa
  • Publication:
    Conference Paper, 14th EADI General Conference
  • Year: 2014


To understand poverty dynamics is needed to analyse not only the factors that are responsible for limiting the welfare of the poor but also the way to reflect this poverty in a reliable measure. There are a plenty of poverty (or deprivation) definition which results on different poverty (or deprivation) measures such as absolute or relative income poverty measure, measure of social deprivation or multidimensional poverty. Though, those which only take into account income/consumption are full of critics for being too narrow. In this regard, poverty is acknowledged as a complex and multidimensional issue so a measure has to try to portray this reality.

Multidimensional poverty index elaborated by United Nations Development Programme (UNPD) includes deprivation in three dimensions such as health, education and living standards in an attempt to capture the whole poverty dynamics, being the most widely used as it covers 109 countries. However, among the main limits of this index appears the non-incorporation of income dimension and the worth or weight given to each indicator, which form the three dimensions. The weights of each indicator are fixed and the same across the countries; they are assigned according to the number of indicators presented in the dimensions.

This paper aim is to contribute to overcome the above-mentioned second limit. To do so, in a first step, Multidimensional poverty index is estimated according to UNPD method, by means of 4,641 face-to-face surveys at household level taken in rural Bangladesh – Chittagong region – in 2012. Bangladesh has a per capita income of 848$ pretending to get 1,300$ in 2021. It is a rural based country (80% of population) and the decline in the national incidence of poverty stood at a mere 9% percent between 1991 and 2008, and about 63 million people still live below poverty line. Second, the multi-criteria technique Analytic Hierarchy Process (AHP) is used to identify the weights of each indicator which from UNPD dimensions according to expert opinion, by mean of 29 face-to-face surveys with researchers working in terms of rural poverty in Bangladesh. Third, significant differences are calculated between the results of the multidimensional poverty index estimating with, on the one hand, fix weights and, on the other hand, expert opinion weights. To increase the accuracy in classifying the households regarding to their poverty level results essential in order to plan and design efficiency policies which allow reducing poverty.

View document