42 Journal of Economic Perspectives Michael Darby(1979, Ch 3)and Laurence Kotlikoff and Lawrence Summers(1981) calculate that the accumulation of households'assets in the United States for the purpose of intergenerational transfers is far more important than that associated with the life cycle. This observation that most people give or receive intergenera tional transfers; a conclusion that supports the ricardian position. franco Modigliani (1988)contests this conclusion, but Laurence Kotlikoff(1988)shows that Modiglianis findings derive from an extremely narrow view of intergenerational transfers Modigliani focuses on bequests at death, and he also does not treat interest earnings on prior bequests as income attributable to intergenerational transfers Some authors accept the idea that intergenerational transfers are important, but argue that the motivation for the transfers matters for the results. Douglas Bernheim Andrei Shleifer and Lawrence Summers(1985) consider the possibility that bequests instead of being driven by altruism, are a strategic device whereby parents induce eir children to behave properly. Some imaginative evidence is presented(involving how often children visit and communicate with their parents) to document the importance of strategic bequests. In this strategic model, if the government redis- tributes income from young to old(by running a deficit or raising social security benefits), the old have no reason to raise transfers to offset fully the government's actions. Instead, the old end up better off at the expense of the young, and aggregate nsumer demand rises. Then, as in the standard approach, real interest rates increase or domestic residents borrow more from abroad One shortcoming of this approach is that it treats the interaction between parents and children as equivalent to the purchases of services on markets. In this setting parents would tend to pay wages to children, rather than using bequests or other forms of intergenerational transfers. These features-as well as the observation that most parents seem to care about their childrens welfare-can be better explained by introducing altruism along with a desire to influence childrens behavior. In this case Ricardian equivalence may or may not obtain. Consider the utility that a parent would allocate to his or her child if there were no difficulty in motivating the child to perform properly. Suppose that the parent can design a credible threat involving bequests that entails the loss of some part of this utility for the child. (Note that if no threats are credible, the whole basis for strategic bequests disappears. If the threat is already large enough to induce the behavior that the parent desires, Ricardian equivalence still holds. For example, if the government runs a budget deficit, the parent provides offsetting transfers to the child, and thereby preserves the child's level of utility, as well as the behavior sought by the parent. On the other hand, the parent may have to allow excess utility to the child to secure a sufficient threat against bad performance. Then a budget deficit enables the parent to reduce the childs utility(as desired), while maintaining or even enhancing the threat that influences behavior. In his case Ricardian equivalence would not hold. Other economists argue that the uncertainty of the time of death makes many bequests unintended, and that such bequests would not respond very much to budget deficits. The imperfection of private annuity markets is usually mentioned to explain