Using data from U.S. commodity flow survey, we show that the historical Union–Confederacy border lowers contemporaneous trade between U.S. states by about 13%. The finding is robust over econometric models, survey waves, or aggregation levels. Including contemporaneous controls, such as network or institutional variables, lowers the estimate only slightly. Historical variables, such as slavery, do not explain the effect. Adding U.S. states unaffected by the Civil War, we argue that the friction is not merely reflecting unmeasured North–South differences. Finally, the border effect is larger for differentiated than for homogeneous goods, stressing the potential role for cultural factors and trust. (JEL F15, N72, N92, Z10)