Abstract <p>The paper analyzes the correlation between the share of food expenditure in Russia and per capita household income using the Engel curve. Overall, taking into account possible measurement anomalies, the behavior of households is consistent classical empirical relationship observed in many countries: as income rises, household food consumption increases, while the share of food expenditure in total consumption expenditure decreases. The paper identified four types of measurement anomalies that can distort the results of the Engel curve estimation. First, when considering long-term macroeconomic dynamics from 2001 to 2023, the classical empirical relationship is robust and statistically significant if the structural shift that occurred during the 2008 crisis is explicitly taken into account. Second, cross-country comparisons of the share of food expenditures and household income for 2019 showed the absence of an abnormally high share of food expenditures in Russia. Cross-country comparisons require particular attention to differences in income levels, taking into account adjustments for imputed rent and eating out in the structure of consumer expenditure. Third, when calculating the share of food expenditures, the Engel curve depends on taking into account the movement of relative prices. This dependence is especially strong in 2001–2007, while after 2010 the contribution from changes in relative prices is weaker. Finally, based on comparisons of the Engel curve parameters obtained using micro- and macrodata, a systematic bias in macroeconomic estimates was revealed. The economic nature of the shifts is likely related to the differentiation of households in terms of overall expenditure and interregional differences in price levels.</p>