A theory of personal wealth transfers is developed which implies that components given to a recipient, such as education and bequests, are perfect substitutes. Therefore, components whose marginal cost rises more rapidly than average reveal wealth elasticities that are smaller than average. For related reasons, time series elasticities are expected to fall short of cross-sectional elasticities. An empirical investigation estimates cross-sectional wealth elasticities of gifts and bequests and time series wealth elasticities of bequest. The estimates are ranked in the order that is anticipated from the theory.