Increasing life expectancy has made working longer both more necessary and more possible, but the relationship between an individual’s survival expectations and his planned retirement age is unclear in the existing literature. This study uses the Health and Retirement Study and an instrumental variables (IV) approach to examine how subjective life expectancy influences planned retirement ages and expectations of working at older ages, and how individuals update those expectations when they receive new information. The estimates in this paper suggest a large and statistically significant relationship between subjective life expectancy and retirement expectations: an individual who is one standard deviation more optimistic about living to age 75 has a greater probability of planning to work fulltime at 62 and 65 by 10 percent to 21 percent, respectively. Respondents who are more optimistic about their survival to age 75 or 85 also expect to work five months longer on average. We also find that increases over time in subjective life expectancy for a given individual are associated with increases in his planned retirement ages and expectations of working at older ages. Finally, actual retirement behavior also increases with subjective life expectancy, but the relationship is somewhat weaker. The results further our understanding of how survival and retirement expectations are “anchored” to the previous generation’s experience and suggest how targeted efforts at increasing knowledge about rising life expectancy may increase the proportion of younger cohorts who decide to work longer.