Many studies in developing countries have found a negative association between breastfeeding duration and child growth. This finding creates a public policy dilemma given the widely accepted benefits of breastfeeding. Plausible causes for the negative association include reverse causality and omitted variables. Using the 2005 Demographic and Health Survey from rural Ethiopia, we find a downward bias in the estimated growth effect of breastfeeding duration before we control for the nutritional adequacy of complementary foods. Once the omitted variable problem is addressed, we find no evidence of adverse growth effects from prolonged breastfeeding. Instead, we find that the effect of inadequate supplementation on child growth is stronger for children that are breast-fed for shorter duration. These results support policies aimed at promoting breastfeeding beyond the first six months of life.