How networked are European venture capitalists at the partner level? Embeddedness, cliques, and structural similarity in a large-scale partner network

Abstract

This paper examines how European venture capitalists connect at the partner level and whether observed ties reflect embeddedness and structural-similarity as predicted by network theory. Using PitchBook data on roughly 41,000 completed European VC deals and approximately 4,000 European-based VC partners, we reconstruct a large person-to-person network from fund-team affiliations, board co-involvement, and portfolio-executive links. The analysis uses network centrality measures, maximal clique participation, local neighborhood overlap, and Monte Carlo SimRank. The empirical design tests two baseline propositions. First, if cohesive participation captures embeddedness and eigenvector centrality captures influence, more embedded actors should be more central. Second, if ties form among structurally proximate actors, connected dyads should exhibit greater similarity than unconnected dyads. The results support both propositions. Clique participation is moderately positively associated with eigenvector centrality. Observed ties also exhibit substantially higher mean Jaccard similarity than randomly sampled non-ties, and the same pattern holds for Monte Carlo SimRank.