I investigate the strong negative relation between recent stock returns and the annuitization of retirement savings using a novel data set with over 100,000 actual payout decisions. After controlling for several standard explanations (e.g., wealth effects), I present evidence supporting naïve...