| Projections by the Social Security Administration suggest that the Old Age and Survivors Insurance trust fund will be depleted in 2035. To shore up the program’s finances, some policy makers propose cutting benefits relative to current law. However, other policy makers express concern that benefit cuts will harm lower-income retirees and instead propose expanding the Social Security system by increasing benefits for large segments of the population, with these increases funded by raising payroll taxes. Because Social Security requires contributions from workers and provides benefits during retirement, theory predicts that expanding the program will crowd out private savings for individuals who are not liquidity constrained. Besides retirement income, Social Security also provides life insurance in the form of benefits paid to the dependent children of deceased workers. Theory predicts that expanding Social Security will also crowd out private life insurance purchases among those who are not liquidity constrained. In this article, we review the existing theory and then explore the empirical evidence for its predictions. Understanding the effect of Social Security on private savings and life insurance holdings is important for a few reasons. First, behavioral responses are an important factor in evaluating how well Social Security meets its objectives. Any potential welfare gains from mandatory saving and risk sharing through Social Security depend on the degree to which households attempt to unwind their Social Security contributions through reductions in private savings and insurance holdings. For instance, Hosseini (2015) shows that the large welfare gains from mandatory annuitization are mostly washed out by the distortions that it causes to the equilibrium price of annuity contracts, as adverse selection causes high mortality individuals to exit the private annuity market. Second, the impact of changes in Social Security rules to maintain fiscal solvency in the face of a demographic shock depends on how individuals respond (Kitao 2014; Bagchi 2016). Finally, if Social Security causes large reductions in personal savings, then growth in the generosity of the Social Security system in recent decades may help to explain the precipitous decline in the aggregate personal saving rate in the United States from near 10 percent in the 1980s to close to zero in more recent years (Parker 1999; Gokhale, Kotlikoff, and Sabelhaus 1996). To motivate our empirical work, we provide a brief review of the predictions of a basic, neoclassical life-cycle model. We use a variant of Yaari’s (1965) classic model to review the theoretical effect of Social Security on private savings by income group. The model features rational individuals who face uninsurable longevity risk and differ by their level of earnings. |