The effects of price on household demand for food and calories in poor countries: are our databases giving reliable estimates?

Food price policy relies heavily on estimated price elasticities of food demand to help balance the nutritional and economic objectives in poor countries. Economists use either unit values (ratios of household expenditure to quantity purchased) or community prices (enumerated from vendors in local markets) as proxies for market prices when estimating price elasticities with household survey data. Biases are believed to result from using unit values, due to measurement error and quality effects, but evidence on this issue is lacking and even less is known about community prices. This article provides an empirical evidence from Vietnam, which suggests that economists should exercise caution when estimating price elasticities from household surveys. A 14-food demand system is estimated alternatively with unit values or community prices, and the elasticity of calories with respect to rice prices is calculated. This elasticity is more than twice as large (-0.54 versus -0.22) when community prices are used rather than unit values. Hence, conclusions about the nutritional effects of rice price increases appear sensitive to data choices made by economists. More generally, this discrepancy suggests that the household survey databases commonly used by economists may not provide reliable estimates.