While the literature finding a decrease in food expenditures at retirement suggests households do not adequately save for retirement, subsequent evidence that nutrient intake is unaffected by retirement has tempered these concerns. We further examine nutrient intake changes at retirement both by analyzing a much wider range of datasets, including longitudinal data, and by improving upon the empirical methodology used in earlier work. Our analysis yields four main results. First, unlike prior work, we find that caloric and nutrient intake fall at retirement in numerous cross-sectional datasets. We can reconcile these contrasting results as being due to well-documented differences and improvements in methodologies used to measure food intake. Second, using longitudinal data, we also find that intake falls at retirement. Third, we show that a food consumption index used in prior work to capture the relationship between permanent income and foods eaten can severely underestimate the impact of retirement on consumption. We show that a minor methodological revision circumvents this bias and that the revised consumption index falls at retirement. Finally, while unemployment reduces the consumption index, we find, in contrast to prior work, that the impact of retirement on the consumption index is larger. Overall, we consistently find that retirement reduces food intake
