AIM: Phylogenetic clustering, the coexistence of evolutionarily related organisms, appears to be common in soil bacteria. This pattern has traditionally been attributed to the habitat‐filtering of bacteria that are able to survive under particular abiotic settings. According to the modern coexistence theory, however, phylogenetic clustering can also arise from biotic interactions such as the competitive exclusion of large clades with low competitive abilities. Here, we used phylogeny‐based methods to discern whether the coexistence of evolutionarily related soil bacteria results from abiotic and/or biotic filtering. LOCATION: Worldwide. METHODS: We performed a Bayesian meta‐analysis based on a literature review (n = 231) to assess whether the net relatedness index (NRI) or the nearest taxon index (NTI), two measures of the phylogenetic relatedness of taxa in local assemblages, deviate from those in randomly configured communities. We then sought the best abiotic (pH, total organic carbon and total nitrogen) and biotic predictors (relative abundance of Proteobacteria, Actinobacteria and Acidobacteria) of NRI and NTI. RESULTS: Phylogenetic clustering is pervasive in soil bacterial communities regardless of the spatial and taxonomic scales (NRI = 2.29; 95% CI 1.43, 3.29; P < 0.001). Clustering is accentuated by productivity; that is, more fertile soils hold communities with more closely related bacteria (estimate = 1.05 0.03, 2.15; P < 0.05). Proteobacterial abundance, which increases with organic carbon enrichment, leads to higher relatedness among coexisting bacteria (estimate = 0.1 0.02, 0.17; P < 0.01) through the competitive exclusion of distantly related deep‐branching clades. MAIN CONCLUSIONS: Our results, together with the dominance of proteobacterial lineages in soils worldwide, suggest that the overrepresentation of this clade underlies the widespread coexistence of phylogenetically related bacteria. These results are consistent with phylogenetic clustering arising via differences in competitive ability as predicted by the coexistence theory. This supports the idea that biotic filtering might have a role in driving the phylogenetic community assembly of soil prokaryotes.