Race statistics and race policy have been intertwined in American history since its founding, starting with the infamous three-fifths clause, continuing with policies based on nineteenth-century race science, the restrictionist immigration at the turn of the century, the Jim Crow regime, and carrying into the civil rights era through such policy concepts as institutional racism, statistical proportionality, disparate impact, and affirmative action. Across this history, the policies and the statistics were about “race,” whether they punished or benefited, were racist or antiracist. But can there be policy that misuses race statistics, that is presented as about race when it should not be? Race statistics are a powerful policy hammer in American history, but not everything is, in fact, a nail. Today the census undercount is argued over as if it is about race; it isn't really. Posing far greater danger, census race categories have worked their way into genomic medicine. The nineteenth-century belief that “race is biological” lingers in the American mind. The use of census categories in genomic medicine risks re-biologizing race. Maybe we should not leave the hammer lying around.