Despite facing some of the same challenges as private insurance markets, little is known about the role of adverse selection in Old-Age Social Security. Using data from the Health and Retirement Study, we perform the unused observables version of the positive correlation test, and find robust evidence that people who expect to live shorter lives both choose smaller annuities – by claiming benefits early – and are less costly to insure, implying adverse selection in the system. Results are consistent when using either subjective expectations or observed longevity. Decomposing the sources of adverse selection we find that health, demographics, occupation and financial information together account for much of the positive correlation between mortality and claiming. IV estimates help to rule out moral hazard.