To understand the magnitude of the problem, this paper, using data from the Health and Retirement Study, specifies how much longer households in each SES quartile would need to work to maintain their pre-retirement standard of living and compares those optimal retirement ages with their planned retirement ages to calculate a retirement gap. It then uses regression analysis to explore whether the gap reflects poor circumstances or poor planning – that is, the extent to which the retirement gap results from health, employment, and marital shocks that occur before the HRS interview but too late for the household to adjust saving (between ages 50 and 58), as opposed to a gap resulting from inadequate foresight. The analysis shows that households in lower-SES quartiles have larger retirement gaps, and this pattern remains true even after controlling for late-career shocks. In short, the most vulnerable have the largest retirement gaps, and these gaps arise from poor planning rather than late-career shocks.