In the wake of the global food crisis of 2007–08 and additional price spikes since then, greater attention has been given to the welfare impact of food price increases in developing countries. The standard approach in this type of analysis, proposed by Deaton (1989), is based on income and expenditure data from household surveys. Given the widespread use of this method, it is important to revisit the assumptions behind it and examine the sensitivity of results to those assumptions. In this paper, we explore the distributional impact of higher maize, rice, and food prices in Ghana and analyze the robustness of those results to changes in several key assumptions. The results suggest that higher maize and rice prices have a relatively modest short-term impact on national poverty but significant effects on specific groups of households. As expected, urban households lose from higher grain prices, but a surprisingly large share of rural households also lose because they are net buyers. The results also suggest that the current policy of protecting domestic rice producers with an import tax does not contribute to national poverty reduction, in spite of the fact that rice growers tend to be poor. If we relax the assumption that households do not respond to the higher prices, the effects are more positive or less negative, but only modestly so. On the other hand, if we relax the assumption that producer and consumer prices rise by the same proportion, and instead assume a constant marketing margin, the results change substantially. Because producer prices now rise by a larger proportion than consumer prices, the impact of higher prices is much more positive. These findings highlight the need for more research on the effect of price spikes on marketing margins.