Given the paucity of quality price data, it is common to rely on "unit value" (average expenditure per unit) as a proxy for price, but this is an imperfect proxy if households respond to price increases by substituting to lower quality goods. This paper draws on survey data that contain both unit value and price to estimate the severity of quality substitution in Indonesia, finding that it is prevalent. The paper next calculates price elasticities that correct for quality substitution, evaluating and ultimately rejecting a commonly used method for calculating price elasticities using only unit value data. Finally, it demonstrates that quality substitution can result in biased price elasticities even when price is perfectly observed.
