This paper tests the Rational Expectations (RE) hypothesis regarding retirement
expectations, controlling for sample selection, reporting biases, and unobserved
heterogeneity. We find that retirement expectations in the Health and Retirement Study
(HRS) are consistent with the RE hypothesis. We also examine how a wide array of
factors, such as wealth, income, health insurance, pensions, and health status influence
retirement expectations formation using panel data from all available waves of the HRS.
We further analyze how new information affects the evolution of retirement expectations
and discover that, on average, individuals correctly anticipate most uncertain events when
planning their retirement, except for some health conditions and economic factors. Our
results have important implications for a wide variety of models in economics that
assume rational behavior.