Dying to Retire: Adverse Selection and Welfare in Social Security

Despite facing some of the same challenges as private insurance markets, little is known about the role of adverse selection in social insurance programs. This paper studies adverse selection in Social Security retirement choices using data from the Health and Retirement Study. We find robust evidence that people who live longer choose larger annuities by delaying the age they first claim benefits, a form of adverse selection. To quantify welfare consequences we develop and estimate a simple model of annuity choice. We exploit variation in longevity, the underlying source of private information, to identify the key structural parameters: the coefficient of relative risk aversion and the discount rate. We estimate that adverse selection reduces social welfare by 2.3-3.5 percent, and increases the costs to the Social Security Trust Fund by 2.1-2.5 percent, relative to the first best allocation. Counterfactual simulations suggest program adjustments could generate both economically significant decreases in costs and small increases in social welfare. We estimate an optimal non-linear accrual rate which would result in welfare gains of 1.4 percent, and cost reductions of 6.1 percent of current program costs.