Economists have recently begun to examine the pattern of financial transfers within families. Both the prevalence and magnitude of such transfers shed light on the motivation behind these and other economic behaviors. Recent papers have focused the discussion of transfers into a debate between altruism and exchange as motivations for the observed behavior. In the context of transfers from parents to children, altruism assumes that a child's utility is an argument in his parent's utility function. Thus, parents transfer resources to children if the child's marginal utility of consumption (appropriately weighted) is greater than the parent's own marginal utility of consumption. One of the predictions of altruism is that the parent will reduce the amount of transfers given to his child if the child's consumption increases. Conversely, an exchange motive views transfers as a payment for services provided by the child to the parent. Unlike the prediction of altruism model, the exchange model does not necessarily predict that the parent will reduce transfers to her child in response to an increase in the child's consumption. The motivation behind transfer behavior has important implications for the effectiveness of government programs. Empirical evidence has been found to support each of the two hypotheses. In this paper the authors provide new empirical evidence on the relationship between the income of the recipients and the likelihood and magnitude of cash transfers using the Asset and Health Dynamics Survey of the Oldest Old (AHEAD). The paper offers strong evidence that respondents are more likely to make transfers and more likely to transfer larger amounts to their less well off children. Contrary to earlier studies, the authors do not contradict an altruistic model of behavior. Furthermore, they present descriptive statistics which cast some doubt on the exchange model and find little if any evidence of an exchange of services for financial compensation.