The widely cited Census Bureau estimates of the number of uninsured people, based on the Current Population Survey, probably overstate the number of uninsured people. This is because of a Medicaid “undercount”: Fewer people report to survey takers that they’re covered by Medicaid than program administrative data show are enrolled. Our study finds that the undercount can be explained by the inability of people to recall their insurance status accurately from the previous year. We suggest that other data sources, such as Census’s American Community Survey, should be studied to determine whether they would provide better estimates of the uninsured.