Estimating Tehsil’s Poverty Ranking in the Province of Punjab from District-Level Survey Data Using Regression Synthetic Estimation

Limited number of observations hamper poverty estimation in smaller areas, rendering national surveys nonrepresentative. National surveys in Pakistan lack adequate sample sizes to generate statistically valid estimates at levels beyond the division and district (i.e., tehsil). This research employs regression synthetic estimation, a small area estimation technique. Poverty in Punjab’s tehsils is estimated using per-adult-equivalent consumption, auxiliary variables derived from national surveys, and census data. Econometric model generates best linear unbiased estimates of per-adult-equivalent consumption at the Tehsil level, providing a spatial view of poverty through the rankings of Tehsils. Empirical results of this research indicate significant inter- and intra-district disparities. This research lifts the veil on the district and shows that peripheral Tehsils are highly impoverished. Findings from this research provide a basis for spatially balanced regional development and support investments in areas outside district headquarters and rural economic development.