Poverty Dynamics in Bangladesh using Pseudo Panel Data

Abstract

There have been numerous studies on measurement and analysis of poverty in Bangladesh over the years. Poverty estimates are mostly based on data of various Household Income and Expenditure Surveys (HIES) carried out by Bangladesh Bureau of Statistics (BBS) which is officially entrusted to conduct the Household Income and Expenditure Survey in Bangladesh. These survey reports provide comprehensive statistical data on various components of income and expenditures at household level, and also their implication to poverty measurement. But these are cross-section data sets and there is a lack of nationally representative panel data at the household level in Bangladesh. Despite an important dimension of poverty there has been hardly systematic study on poverty dynamics in Bangladesh based on quantitative evidence due to lack of such panel data sets tracking the same households over time. Ideally, monitoring of poverty requires repeat surveys of the same households – known as the panel survey – so that individual household characteristics can be controlled for in evaluating poverty incidence and inequality overtime. Bangladesh Bureau of Statistics, however, never attempted at formulating panel survey. From the real panel data we can explore the life trajectories of the panel households as same households of the samples are followed over-time. We can draw valuable findings from long-run analyses that can hardly be obtainable in short-run analyses using cross-section data sets. Pseudo panel data are being used extensively in the current economic literatures particularly when there is dearth of true panel data in the developing countries. Using Pseudo panel data are now well established as a substitute for the unavailable true panel data. The study has made an attempt to construct a pseudo panel data set using four rounds of nationally representative HIES data (1991/92, 1995/96, 2000 and 2005) collected by BBS. Pseudo panel data has been constructed with the formation of cohorts by grouping households on the basis of shared characteristics to explore the dynamics of poverty in Bangladesh. It is observed that the momentum of escaping from poverty of the day labor group, particularly the day labor engaged in agricultural activities is very low compared to other types of occupation groups like self employment in agriculture or non-agriculture, or service holders between 1991 and 2005. The study also explores poverty dynamics scenario among the various core household groups ‘never poor’, ‘movers’, ‘fallers’ and ‘chronic poor’ over the same period. Various econometric techniques have been applied while exploring determinants of poverty using pseudo panel data.

M. Abu Eusuf

PhD Candidate, IDPM, University of Manchester

E-mail: eusuf101@yahoo.com

See for more information:

Roy Huijsmans