Monitoring Poverty in a Data Deprived Environment: The Case of Lebanon

This paper is motivated by the dearth of statistical capacity in the Middle East North Africa region and the unprecedented economic collapse in Lebanon. It proposes and illustrates a data augmentation approach to conduct poverty analysis in the absence of traditional sources of information on income distribution. Our approach shows that it is possible to exploit alternative data sources to conduct the much-needed poverty analysis. Building on available data augmentation techniques, we first recover the entire income distribution from the available interval data. Then we account for nonresponse and estimate the bounds of the set of admissible cumulative distributions of income. Finally, we analyze poverty dynamics using first-order dominance tests on the bounds of admissible cumulative distributions set. To illustrate the importance of the proposed approach, we apply this methodology to Lebanese data, provide a picture of poverty dynamics, and provide insights into the politico-economic dynamics preceding the economic collapse.