Examination of household (micro) data on the elderly has generally con
cluded that they do not dissave significantly, whereas estimates using
aggregate (macro) data have shown that the presence of a large elderl
y population leads to a lower saving rate. This paper shows that if in
teractions between generations are important, one would not expect the
se estimates to be the same. The paper presents new evidence that bequ
ests are plausibly the source of this discrepancy. Examination of data
from young households that have received or expect bequests confirms
that bequests are indeed an important factor determining the saving of
the young.