Consumption is a better measure of well-being than income particularly among the retired because of the ability to spend out of assets. Yet, because of very limited sources of data on consumption, income is almost universally used to measure well-being for policy and for economic analysis. The life-cycle model is the leading model to explain consumption and saving decisions, and it makes strong predictions about how consumption should vary over the life-cycle. But because of a lack of panel data on consumption in the U.S., many analyses have been based on synthetic panels from the Consumer Expenditure Survey or on partial measures of consumption such as food consumption in the Panel Study of Income Dynamics. This project will use the Consumption and Activities Mail Survey, which is a three-wave panel on spending based on 5000 randomly chosen HRS households, to study consumption paths from the immediate pre-retirement years to advanced old age. The project will compare welfare indicators (such as poverty) based on consumption with those based on income; find the response of consumption and saving to anticipated events such as retirement and unanticipated events such as health shocks or stock market gains or losses; and estimate models of consumption. These models will incorporate returns-to-scale in consumption, mortality risk, and health-care expenditure risk.