Social Security Increases Wealth Inequality

Social Security differentially disenfranchises the children of low and middle income families from receiving inheritances.

In Simulating the Transmission of Wealth Inequality Via Bequests (NBER Working Paper No. 7183), Jagadeesh Gokhale, Laurence Kotlikoff, James Sefton, and Martin Weale find two things that run contrary to popular belief. First, most U.S. inheritances may be unintended. Second, were it not for Social Security, inheritances would probably mitigate, rather than exacerbate, U.S. wealth inequality among married couples reaching retirement.

Absent social security, unintended inheritances can reduce wealth inequality because their receipt is random, depending on when parents pass away. But this wealth equalizing aspect of inheritances -- that the parents of high income children may die late and those of low income children may die early -- is undone by Social Security. The reason is that Social Security differentially disenfranchises the children of low and middle income families from receiving inheritances. It does so by annuitizing a much larger share of the economic resources of those parents than of rich ones.

In addition to Social Security, other factors contribute to wealth inequality. The most important of these is inequality in lifetime earnings, because high earners will save more for retirement than low earners. A second critical factor is the degrees of assortative mating (choosing a spouse based on earnings potential), since this process exacerbates inequality in total household earnings.

The authors' analysis assumes a life span of 88 years, divided into distinct economic and socio-demographic phases. The calculations closely approximate the degree of inequality in the actual U.S. data. For example, in the authors' calculations the richest one percent of retirement households own 32.8 percent of total wealth, compared to an actual figure of 30.4 percent according to the Survey of Consumer Finances (SCF).

All told, the factors that the authors examine through their simulation model are capable of reproducing much of the observed wealth inequality in the United States and closely match the SCF data.

-- Lester A. Picker