Aggregate data and income elasticity of demand: the example of medicines in Brazil

Purpose This paper aims to demonstrate how the level of data aggregation affects the estimation of the income elasticity of demand and potentially influences the distinction between luxury and necessary goods. We explore the estimation of the income elasticity of outpatient medicine demand in Brazil. Medicines may compete with expenditures on essential goods, which turns them into a frequent target of public policies worldwide. Design/methodology/approach We used a single data set (the Brazilian Household Budget Survey – BHBS, 2017–2018) with varying levels of data aggregation and an identical method of estimation to assess how the level of aggregation affects elasticity coefficient estimations. Findings The elasticity estimated based on per person microdata significantly differed from that estimated on a per capita basis by primary sampling unit and by State, but was similar to per capita estimates by household. Originality/value Although many academic studies still use aggregate data to estimate associations between variables that represent individual characteristics or choices, this procedure is incorrect and can lead to inconsistent results. Data aggregation level matters when it comes to understand and estimate demand elasticity. Elasticities based on aggregate data provide overestimations unfit both for crafting public policies and for projecting future demand.