Simulating the Transmission of Wealth Inequality

Abstract

The role of bequests in propagating wealth inequality has long interested economists, policymakers, and social commentators. Josiah Wedgwood’s (1929) study of wealthy Britons indicated that most had received large inheritances and suggested that one-third owed their position in the wealth distribution entirely to inheritance. These findings and those of J. E. Meade (1966), C. D. Harbury and D. M. W. N. Hitchens (1979), and Paul L. Menchik (1979) support the public’s general view that restricting inheritances by taxing estates or inheritances or by forcing annuitization would lead to a more equal wealth distribution. Would wealth inequality actually be reduced if inheritances and, for that matter, all private inter vivos transfers were eliminated? The answer is not obvious. Private wealth holdings would, in this case, be traced to precautionary and retirement saving, with the distribution of wealth determined by the distributions of aftertax labor earnings, rates of return, demographics, and saving preferences. Household labor income is distributed very unequally because of differences in genetic endowments, educational opportunities, parental care, health, labor– leisure preferences, assortative mating, and a host of other factors. Rates of return earned on saving also vary widely across households for systematic and nonsystematic reasons. Also, as recently documented by Steven F. Venti and David A. Wise (2000), there is a great deal of heterogeneity with respect to saving behavior. Thus, a high degree of wealth inequality would exist in the absence of bequests and inter vivos gifts. Adding them back into the mix could actually reduce overall wealth inequality if they were made to children with relatively low earnings, low rates of return, or poor saving discipline or were made primarily to children whose parents died young. Clearly, understanding the precise role that bequests and gifts play in wealth inequality requires building a fairly elaborate model that can control for different factors and their interactions. The model we co-developed in Gokhale et al. (2001) to study U.S. wealth inequality, which we extend here, represents a step in this direction. The model features 88 overlapping generations. It incorporates marriage, fertility patterns, random death, heterogeneous skill endowments and rates of return, assortative mating based on skills, skill inheritability, progressive income taxation, and wealth annuitization via Social Security. Given the strong evidence against intergenerational altruism reported in Kotlikoff (2002) and the dictates of tractability, we modeled bequests as arising solely from imperfect annuitization. The model generates a realistic ratio of aggregate wealth to aggregate labor income, a realistic flow of bequests relative to the stock of wealth, and a realistic distribution of wealth at retirement, including the share of wealth held by those in the top tail of the distribution. Bequests play a limited role in influencing wealth inequality, the major determinant of which is skill (earnings) differences. Interestingly, bequests serve to equalize the distribution of wealth because, when children inherit, wealth is determined by the random date of parent’s death. In contrast, Social Security plays a disequalizing role. As stressed by Martin S. Feldstein (1976), Social Security annuitizes a much larger share of the assets of the poor than of the rich, leaving them with relatively little fungible wealth.