The (dynamic) effect of retirement on food purchases

Abstract

We investigate the potential dynamics in the effect of retirement, on food-at-home purchases and dietary quality in Denmark. We combine Home-scan data with nutritional information and administrative data on retirement, income and health status. The panel contains 497 retiring and 1,471 control households. We estimate the effect of retirement up to 10 years after the date of retirement by using Fixed Effects with health and wealth proxies, as well as Fixed Effects IV with the ages eligible for retirement as instruments to control for the potential endogeneity of retirement. Based on the Fixed effects results we find that overall dietary quality increase slightly at retirement, but find only minor and mostly insignificant changes in the individual components of the diet. The effects are found to be of the same magnitude, but insignificant in the FE-IV estimation. Hence, there seem to be a small increase in dietary health upon retirement in Denmark. The results for food expenditure and energy consumption are ambiguous. Based on the FE with proxy variables we find indications of long-run adjustments in food expenditures, while energy consumption is immediately affected by retirement, but has no further adjustment. Both effects are insignificant in the FE-IV estimation. The very small changes observed, suggest that dietary behaviour might be governed by habitual behaviour and might also be due to the high income replacement rate at retirement in Denmark.