Monitoring poverty in a data deprived environment: The case of Lebanon

Abstract

This paper is motivated by the dearth of statistical capacity in the Middle East and North Africa region and the unprecedented economic collapse in Lebanon. We expand and apply a novel data augmentation technique to conduct poverty analysis when the usual data sources on income distribution are limited or unavailable. Building on available data augmentation techniques, we recover the continuous income distribution from the available data when the income variable takes the form of intervals. We expand existing techniques to derive dominance conditions for interval income data accounting for non-response. This extension allows us to run robustness checks of our empirical results by estimating the bounds of the set of admissible cumulative distribution functions. Our empirical application then analyzes poverty dynamics using first-order dominance tests on the bounds of admissible cumulative distribution functions sets and shows the importance of the proposed approach using Lebanese data. The empirical application provides a picture of poverty dynamics and insights into the politico-economic dynamics preceding and following the economic collapse. More generally, we show that the development analyst can exploit alternative data sources to conduct the much-needed poverty analysis.