Consumption Subaggregates Should Not Be Used to Measure Poverty

Abstract

Frequent measurement of poverty is challenging because measurement often relies on complex and expensive expenditure surveys that try to measure expenditures on a comprehensive consumption aggregate. This paper investigates the use of consumption “subaggregates” instead. The use of consumption subaggregates is theoretically justified if and only if all Engel curves are linear for any realization of prices. This is very stringent. However, it may be possible to empirically identify certain goods that happen to have linear Engel curves given prevailing prices, and when the effect of price changes is small, such a subaggregate might work in practice. The paper constructs such linear subaggregates using data from Rwanda, Tanzania, and Uganda. The findings show that using subaggregates is ill advised in practice as well as in theory. This also raises questions about the consistency of the poverty tracking efforts currently applied across countries, since obtaining exhaustive consumption measures remains an unmet challenge.

Fewer Questions, More Answers: Truncated Early Stopping for Proxy Means Testing

“…Focusing on settings with cost and operational constraints on data collection, Pape and Mistiaen [2018] proposed an effective imputation approach to estimate the population distribution of consumption rather than a full LSMS-style survey. A similar focus on a reduced length survey is Christiaensen et al [2021], who investigated the use of components of consumption to estimate household values but find both theoretical and empirical problems with this approach.…”

Using Survey-to-Survey Imputation to Fill Poverty Data Gaps at a Low Cost: Evidence from a Randomized Survey Experiment

“…In this respect, only two studies exist, and they offer inconclusive evidence. While Christiaensen et al (2022) suggest that using consumption sub-aggregates for poverty imputation only works under certain stringent conditions, analyze 14 surveys from various countries and demonstrate that adding household utility expenditures to a basic imputation model with household demographic and employment attributes can produce accurate poverty predictions -consistently within the 95 percent confidence internal, and often within one standard error, of the observed "true" poverty rate. 2 Finally, the last issue motivating our work is that existing studies that "validate" imputed poverty estimates were implemented in artificial settings.…”

Households in Transit: COVID-19 and the Changing Measurement of Welfare

“…We also build on previous work on imputing values of unobserved expenses or incomes, such as the literature on rent imputation for owner-occupied housing (Balcázar et al 2017). We build on previous work showing theoretical problems with the consumption aggregate, including showing that partial consumption aggregates which do not collect information on all goods only satisfy desirable theoretical properties if the associated Engel curves are all linear (Christiaensen, Ligon, and Sohnesen 2021). We offer a general method that can be applied to other contexts, settings, and datasets.…”