Hybrid Survey to Improve the Reliability of Poverty Statistics in a Cost-Effective Manner

Abstract

The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.

Is Predicted Data a Viable Alternative to Real Data?

“…However, the standard errors for the national poverty rate derived from a small sub-sample of 640 households are 2.9 and 2.4 percentage points (see Table 5 of Ahmed et al (2014)), respectively, for the sample direct and prediction estimators, which corresponds to a 32 percent (≈ 1 − (2.4/2.9) 2 ) reduction in variance. Without further information, however, it is hard to determine how much of this cost-precision trade-off can be attributed to the fact that their double sampling estimator substituted predicted data for real data.…”

“…To further elucidate this point, consider a study by Ahmed et al (2014), which implies considerable gains from using a prediction estimator in an application to poverty measurement in Bangladesh. Unfortunately, their approach does not fit perfectly into our analytical framework such that we are unable to compute the values for (w c , w h ) that would apply to their data.…”

“…In Bangladesh, Ahmed et al (2014) adopt the assumption that five enumerators have to stay for two weeks to complete the survey in one primary sampling unit (PSU) to estimate the cost of conducting Household Income Expenditure Survey in Bangladesh. They further assume that two persons engage in collecting consumption data and three persons collect non-consumption data for two weeks in one PSU.…”

“…Assessments of the cost-precision trade-offs however are still limited. Ahmed et al (2014) denotes an exception which provides a SWIFT-like application of double sampling to poverty estimation in Bangladesh, where a variety of different data collection scenarios are being considered. Their simulation results indicate that substantial cost savings could be achieved with a moderate loss in precision.…”

Mapping Poverty in Afghanistan

“…However, this requires dividing the survey data by province, so as a result fewer observations are available for each province model. As a rule of thumb, one would like to have no less than 300 observations for each model (Ahmed et al, 2014), which was satisfied for Kabul and Herat.…”