The world population is ageing, and as lifestyle related illnesses and multi-morbidity increase with ageing, prevention with modifiable factors is becoming an increasingly important issue. In the current paper, we revisit the retirement (food) consumption puzzle in a welfare state context and examine how retirement affects food expenditure and energy intake as well as the dietary healthiness of elderly individuals with different marital and labor market status. We use a unique household panel data set covering the period from 2007–2017 with monthly food purchases from a sample of households above 55 years of age. We supplement the purchase data with official register data on income, health-contact days, unemployment, early voluntary and old age retirement at the individual level. The structure of the data set facilitates the identification of any associations between retirement and dietary behavior while controlling for important confounding factors such as the number of health-contact days as a proxy for health shocks and income as a proxy for wealth shocks. We compare the results with the results from estimations using instruments for the retirement decision. We do not find any evidence to support a retirement (food) consumption puzzle. Generally, the healthiness of diets increases at the point of retirement, and even more so for those who had the most unhealthy diets before retirement. Since the diets of individuals outside the labor market are less healthy both before and after retirement compared to other groups, the focus of dietary policy should be directed at this group.