Using local expert knowledge to measure prices: Evidence from a survey experiment in Vietnam

Many countries lack spatially disaggregated consumer price data. Yet these data are needed to estimate real inequality and spatial patterns of poverty, especially for poor countries where weak infrastructure and high transport costs create big price variation over space. We experimented in Vietnam with a new way of obtaining disaggregated price data, using local expert knowledge to derive the mean and variance for prices of 64 consumer items in over 1000 communities. We used photographs of the specified items to ensure comparability of the reported prices. These prices are used to calculate regional cost-of-living indexes, which provide a good approximation to benchmark multilateral price indexes that are calculated from data obtained from traditional market price surveys. In comparison, two widely used no-price methods, based on using food Engel curves to derive deflators and based on using unit values (survey group expenditures over group quantity) are very poor proxy indicators of prices and of the cost-of-living and would distort estimates of real inequality and the spatial pattern of poverty. Prices from local expert informants also exhibit a basic spatial feature of prices – the Alchian-Allen effect or ‘shipping the good apples out’ – in much the same way as do prices from the traditional survey approach. This effect is one reason why unit values are a bad proxy for prices and this effect should become more important as food systems commercialize. Using expert knowledge to measure local prices is a low-cost and feasible approach that could be adopted more widely in developing countries.