A quality-index of poverty measures

The multitude of available poverty measures can confuse a policy maker who wants to evaluate a poverty-reduction policy. We proposes a rule for ranking poverty measures by use of the food-gap, calculated as the cost-difference between a household’s normative food basket, derived from a healthy diet, and the actually chosen food basket. The rationale for this indicator is based on the fact, that (1) basic food needs reflect an ultimate necessity, (2) food expenditure is highly divisibility, thus allowing for efficient marginal substitution between competing necessities when the household’s economic hardship increases. For these reasons we believe this to be an objective indicator for the sacrifice in the standard of living of a family under economic stress. A household is identified as ‘truly’ poor or non-poor by a given poverty measure if the diagnoses coincide and vice versa. The ranking is obtained by a gain-function, which adds up congruent and deducts contradicting outcomes for each poverty measure. We calculate four types of gain-functions –of headcounts, food-gaps, FGT-like powered food-gaps and an augmented version of the latter. The poverty measures include expenditure-based, income-based, relative, absolute, mixed measures and a multidimensional measure of social deprivation. The most qualitative measure is found to be Ravallion’s Food Energy Intake and Share measure, though it suffers from a possible bias, since it includes the food-norm in its design. The 60%-median income measure from all sources ranks highest among the unbiased measures. The absolute poverty measure yields the worst performance.