Deflation by expenditure components: a harmless adjustment?

We investigate the effect that seemingly minor features of the implementation of cost-of-living adjustments have on the distribution of household expenditures, by developing an analytical framework that is consistent with standard consumer theory, and mindful of data limitations faced by practitioners. The main result is at odds with common sense: even when multiple price indices are available (e.g., a food and a non-food CPI), it turns out that using a single price index (e.g., the total CPI), to adjust the consumption aggregate is recommended. The practice of adjusting sub-components of consumption separately (food with a food index, non-food with a non-food index) can lead to a systematic bias in the welfare measure, and consequently in poverty and inequality measures. Using Iran’s 2019 Household Expenditures and Income Survey, we find that the bias manifests as a systematic underestimation of urban poverty, and overestimation of rural poverty.