Prior research finds food expenditure decreases at retirement, which suggests households are inadequately saving. In contrast, other evidence shows that direct measures of food intake are unaffected by exiting the labor force. Using a wide array of data sources and methodologies, we find that food intake falls at retirement, including: declines in caloric and nutrient intake in cross-sectional datasets spanning 40 years, a decrease in caloric intake using longitudinal data, and a drop in a food intake index that relates diet composition to permanent income. We discuss the implications of these results in the context of life-cycle models of consumption. (JEL D15, G51, I12, J26)