After a decade of research, the effect of inequality on long-run economic growth remains unresolved, in part because researchers have treated omitted variable bias as an estimation problem rather than a deeper question of causality. In this article we argue that the key omitted variable is the...