The authors argue that endogeneity of transportation costs needs to be taken into account when estimating the effect of distance on trade. Otherwise, the estimates of the distance effect may be biased and inconsistent. Endogenous transportation can introduce slope heterogeneity and spatial correlation. Both issues can be accommodated with the help of Pesaran's cross-correlated effects mean-group (CCEMG) estimator. After applying this methodology, the authors uncover significant compression of the distance effect on trade starting from the middle of the 1990s. The trade-reducing effect of long distances becomes statistically indistinguishable from the effect of moderate distances. This compression is not present in the traditional fixed effects estimates. The authors hypothesize that such pattern may be reconciled by changes in shipping technology that disproportionately reduce transportation costs over long distances.