Tropical mountains harbor exceptionally high biodiversity, which is in part due to the marked elevational stratification of tropical biotas. However, the factors that influence the evolution of elevational distributions remain uncertain. I used a database of sister species of tropical montane birds from 41 families and three regions—the Neotropics, the Himalayas, and New Guinea—to test whether patterns of elevational divergence were consistent with (1) a stochastic process, (2) ecological sorting of elevational divergence that occurred in allopatry, or (3) elevational divergence driven by competitive interactions upon secondary contact. The stochastic and ecological sorting hypotheses predict that increased elevational divergence in sympatric sister species is explained by their greater evolutionary age, whereas the competitive interactions hypothesis predicts that elevational divergence is explained by geographical overlap. I found that genetic distances were unrelated to elevational divergence and that allopatric sister species occupied similar elevational distributions regardless of genetic distance in each region. Instead, sympatry was the only significant predictor of elevational divergence; regardless of evolutionary age, sympatric sister species had greater elevational divergence than allopatric sister species in each region, as predicted by the competitive interactions hypothesis. Importantly, this pattern occurred in all three geographic regions, which suggests that competition-driven elevational divergence upon secondary contact is a general process of community assembly in tropical montane avifaunas.