| Abstract: | We distinguish between bequests that are planned as part of a lifetime optimization process and those that are the unplanned result of errors in forecasting the date of death. Given the initial endowment we expect a positive relation between the bequest and the planning horizon, and a negative relation between unexpectedly long life and the bequest. A unique data set on wealthy decedents and their parents provides measures of expected horizons based on parents' longevity. Both predictions are confirmed, and there is an especially large and significant positive relation between the bequest and the length of the planning horizon. |
