To begin with some very informal evidence, Kennickell, Starr-McCluer, and Sun-den (1995) report some results from a “focus group” session on saving motivations that was convened as part of the preliminary work in designing the questions for the 1995 SCF.6 The eight members of the group were all wealthy individuals, mostly in their 50s. Participants were asked “Thinking about your reasons for saving, what sorts of reasons are most important to you?” In the entire course of a three hour conversation of saving behavior, however, providing a bequest was not mentioned once as a reason for saving.
|Most||One of the 5||Number|
|Richest 1 Percent||.02||.04||652|
A group of eight individuals is obviously too small a sample to convincingly demonstrate the general absence of a bequest motive among the wealthy. Somewhat more persuasive evidence is provided in the results of survey questions on the 1992 SCF. Respondents were asked to list their five most important reasons for saving. As shown in Table 1, only three percent of the general population, and two percent of the wealthy households, indicated that providing an inheritance was the most important reason to save. Furthermore, only 5 percent of the total population and 4 percent of the wealthy households indicated that providing an inheritance was among their top 5 reasons for saving. (The differences between the wealthy households and the general population are not statistically significant here.)
Another obvious test of the model is to see whether the childless elderly tend to dissave more than those with children. This hypothesis has been tested using population-representative data; Hurd (1986) found that in the population as a whole, there is no tendency for elderly with children to decumulate faster than those without. Unfortunately, even when the data from the four SCFs are combined, the number of childless, elderly, wealthy households is too small to permit reliable estimation of age profiles of wealth (only about ten percent of elderly couples are childless).
Another option is to consider what childless elderly people say about their saving and spending behavior. Respondents to the 1992 and 1995 SCFs were asked whether their spending was greater than, equal to, or less than their income over the past year, and how spending usually compared with income. The results are presented in Table 2.10 The childless elderly were less likely to say that they dissave than those with children, by this crude measure, either as a general rule or in the current year in particular. Of course, it is possible that some of the “spending” of the elderly with children consists of inter vivos transfers to those children. The real problem for the Life Cycle model is the testimony of the childless, wealthy elderly, essentially none of whom say that their spending exceeds their income. This is all the more impressive given the comparatively small fraction of their income that is annuitized.