Multidimensional Poverty Analysis: assessing poverty patterns among different demographic groups of India

Abstract

The concept of Multidimensional Poverty Index (MPI), which is relied upon the idea that many intertwined aspects of poverty might not be always possible to be measured in terms of fixed prices, has gained considerable familiarity over the last decades and there have been various experimentations to determine the aspects which should be most imperative to review in order to determine who are poor and who are not. In almost all studies, education, health and standard of living have been singled out as the three obligatory aspects that cannot be avoided if we are rightly trying to estimate poverty and the Alkire-Foster counting approach is the most widely acknowledged and implemented method for constructing Multidimensional Poverty Index. In this paper we have extended the idea, by using the latest 2021 Alkire and Kanagaratnam proposition, for 2015 Indian National Family Health Survey Data and we have observed significant changes (compared to the previous case studies for 2015 NFHS survey data where any older MPI computing methods were used) in the counts and ratios across India’s different states, caste groups, religious groups. The paper also looks for a method to extend the MPI counting from household level to individual level. The exploratory analysis wrt almost all possible diverse indicators of deprivation across different religions, caste groups, states and other demographic groups, is another key research aspect of this paper with the motivation to make the results easily interpretable and reusable even outside the scope of MPI construction.  It’s certain that the poverty level in India is not same for different demographic groups. This paper implements statistical tools such as Multiple Regression and Principal Component Analysis to assess the significance of the findings from the exploratory studies.