Rapidly rising prices for staples have reawakened policy concerns about food security in Asian countries. While these concerns are often expressed through national self-sufficiency policies, considerable attention is also paid to micro-level food security. At the micro level, the question of whether the poor pay more for food remains contested in the economics and geography literature, after more than four decades of study. In this literature, two broad approaches have been used; store surveys to compare prices of identical foods in rich and poor neighborhoods, and household surveys to compare unit values (the ratio of expenditure to quantity) across rich and poor households. While household surveys suggest that the poor pay higher prices, store surveys typically find food prices varying with store type and location but not with neighborhood income, given location. In other words, there is no price penalty for the poor. But store surveys are mostly used in developed countries, with only one study from a developing country (for 14 foods in one region of Brazil). In this paper we use especially collected food price data from metropolitan areas of Vietnam to test whether the urban poor face higher food prices. We also link the price surveys to a household survey to examine whether both types of data yield the same answer to the question of whether the urban poor pay more for food.
Data/Methods:
In 2010 the lead author designed a price survey for 45 foods in 210 urban locations that was fielded by the General Statistics Office (GSO) of Vietnam. The survey covered both branded and unbranded foods, and items that are typically unquantified in surveys, such as street meals. To maintain consistency of item specification across areas, enumerators used detailed photographs of the items whose price was required. A quota sample of stores, registered outdoor markets and unregistered street markets was used, within districts that were stratified according to neighborhood income. We use these data to test whether food prices are systematically higher in poorer districts of these metropolitan areas. We also use unit values from a household survey carried out at the same time to test whether any apparent price penalty for the poor is confirmed with both price survey data and unit value data.
Expected Results:
For each food, mean prices for poor, middle income and rich districts will be reported, along with t-tests of equality. We will also report an analysis of covariance, to see if the income strata effect on each food’s price is more important than the metropolitan location effect. The relative prices of all 45 foods will then be combined with household budget data to calculate a Törnqvist food price index for each metropolitan area, and a test will be conducted of whether the spatial pattern of relative prices differs when using prices just from the poor stratum compared with using prices just from the rich stratum. Finally, unit value evidence for whether the poor pay more will be compared with the results of the price surveys.