Essays on poverty in Sri Lanka

<p>Sri Lanka is a lower-middle income country in South Asia. The incidence of income poverty in the country has been in sharp decline over the last two decades. However, there are on-going debates about how poverty is best measured. Using a binary measure of poverty fails to identify the non-poor that are vulnerable to poverty as well as those that are deeply entrenched in poverty. Moreover, it is widely accepted that poverty should be defined and measured using indicators other than income. This study attempts to answer four research questions. First, using income, what is the prevalence of poverty, extreme poverty, vulnerability to poverty, and non-poverty and how have their prevalence changed through time? Second, what is the incidence of multidimensional poverty for Sri Lanka and, using population surveys, what weights should be attached to the dimensions of a Multidimensional Poverty Index (MPI)? Third, what are the determinants of income and multidimensional poverty in Sri Lanka and how do they differ? Fourth, given the paucity of data, what are the missing dimensions of multidimensional poverty in Sri Lanka that should be included in future measures?</p> <p>The study employs data from Sri Lanka's Household Income and Expenditure Surveys (HIES). Primary data were also collected through a Discrete Choice Experiment (DCE) to determine people's preferences over weights attached to the different dimensions of a MPI. Quantitative data are analysed using a variety of econometric techniques. Qualitative data were also collected through key informant interviews to identify the most important missing dimensions of the MPI devised by the paper. Findings suggest that while the incidence of poverty and severe poverty is low in Sri Lanka, a much larger proportion of households are vulnerable to poverty, in the sense that they only just above the poverty line threshold. An MPI is constructed that incorporates information on the following three dimensions of wellbeing: health; education; and living standards. Findings from the DCE suggest that the health dimension should receive the highest weight followed by education and living standards. MPI rankings of districts and provinces differ across equal weights and DCE weights applied to the index.</p> <p>Other findings suggest that income poverty and multidimensional poverty share many of the same correlates. Age of the household head, education, marital status, employment, access to agricultural land and the receipt of remittances are important across the two poverty measures. However, there are important differences with respect to the signs of the coefficients on household size and ethnicity variables. Further, the magnitude of the associations of independent variables with poverty vary considerably across the two measures as well as their arbitrary poverty line thresholds.  Finally, the study proposes an augmented MPI which includes information on food security, safety and security and access to basic services.</p>