We study the role of longevity and rising income inequality in the growth of wealth inequality in the United States. A large body of literature documents a rise in income inequality and attributes growing wealth inequality in the U.S. to rising income inequality. Hubmer et al. (2021) argue that maintaining redistribution in the tax system could have prevented the rise in wealth inequality even in the presence of high income inequality. At the same time, the U.S. experienced a colossal increase in life expectancy, especially in old-age longevity. Between 1970 and 2015, life expectancy at 65 has improved from slightly above 14 years to nearly 19 years. Through the lens of any standard overlapping generations model, this rise in longevity can translate to an increase in wealth inequality due to two mechanisms. First, since individuals expect to live longer, a behavioral effect involves higher wealth accumulation at the peak of the life cycle for each subsequent birth cohort. Second, in the first stage of the demographic transition, a composition effect occurs due to a rising share of individuals close to the peak of wealth accumulation rises.

Our main contribution to the literature is the finding that, indeed, demography matters for wealth inequality. We use our setup, a rich general equilibrium overlapping generations model, to perform counterfactual simulations, where we set mortality risk, income mechanisms, and taxes at levels from the 1950s. These counterfactual scenarios reveal that the rise in longevity sizably contributed to the rise of wealth inequality. In fact, through our model's lens, longevity's contribution to wealth inequality is as important as income inequality. Further, our simulations show that the behavioral channel driven by consumption-savings responses to longer lifespans is quantitatively more relevant than the changes in the demographic structure. Given the demographic evolution, the role of the tax system is minor. We use demographic projections to study the future evolution of wealth inequality. We show that the forces currently at play will continue to operate for another half century, and the current tax system impact will be minor.

We replicate the features of the U.S. economy in the 1930s in the initial steady state and study the transition to the new steady state. We introduce changes in longevity by adjusting survival probability in line with the demographic data and projections for the U.S. Regarding income changes, we use data on the share of college-educated individuals and the skill premium over the available period. We estimate age profile and income risk from PSID. The estimated income risk varies along the transition path for subsequent birth cohorts. Regarding redistribution, all tax rates and income tax progressivity are taken from the data. The model generates income and wealth inequality dynamics that closely track the magnitudes in the data: a decline in wealth inequality Gini coefficient until the late 1970s and a subsequent increase of 5 points since then.