For many research problems in developing countries, some information on prices faced by households is required for the analysis, but these prices are not readily available from household surveys, nor is it straightforward to observe them, especially if subsistence onsumption is a substantial part of consumption. Furthermore, quantities consumed and produced are often in local units presenting further problems for the analysis. Building on Deaton’s (1987) seminal work, we provide an econometric approach to estimate prices and quantity conversion factors from household expenditure data.. We use panel data from rural Ethiopia to illustrate the approach and to investigate the potential quality bias in the estimation of the prices. In an application we show that the conclusions about poverty
changes over time are significantly affected by using less appropriate strategies to convert local units and to value subsistence consumption.