Accumulating evidence suggests that agricultural production could be greatly affected by climate change, but there remains little quantitative understanding of how these agricultural impacts would affect economic livelihoods in poor countries. Here we consider three scenarios of agricultural impacts of climate change by 2030 (impacts resulting in low, medium, or high productivity) and evaluate the resulting changes in global commodity prices, national economic welfare, and the incidence of poverty in a set of 15 developing countries. Although the small price changes under the medium scenario are consistent with previous findings, we find the potential for much larger food price changes than reported in recent studies which have largely focused on the most likely outcomes. In our low productivity scenario, prices for major staples rise 10-60% by 2030. The poverty impacts of these price changes depend as much on where impoverished households earn their income as on the agricultural impacts themselves, with poverty rates in some non-agricultural household groups rising by 20-50% in parts of Africa and Asia under these price changes, and falling by equal amounts for agriculture-specialized households elsewhere in Asia and Latin America. The potential for such large distributional effects within and across countries emphasizes the importance of looking beyond central case climate shocks and beyond a simple focus on yields - or highly aggregated poverty impacts.