In a recent paper in this journal, Bloom, Canning and Graham (2003) model the effect of changes in longevity on individual savings. They proceed to present empirical findings about the relation between longevity and "aggregate savings." There is a missing link between their empirics and theory: the changes in the populations age density distribution due to increased longevity. This note provides such an aggregation analysis within a simple model with uncertain survival, endogenous life-cycle consumption and retirement age. It is shown that, with continuous annuitization, an increase in expected longevity raises aggregate steady-state savings. The magnitude of this effect depends on the economy's age-specific distribution and on the elasticity of optimum retirement to changes in longevity. Copyright The editors of the "Scandinavian Journal of Economics" 2006 .
