Venture Capital Communities

Abstract

While VCs can choose from thousands of potential syndicate partners, many prefer to co-syndicate with partners drawn from small groups of preferred partners, whom we term "VC communities." Using computational methods from the physical sciences, we identify communities from three decades of syndication data and characterize their number, size, composition, and economic activities. Communities comprise VCs similar in age, connectedness, and functional style and undifferentiated spatial locations. Machine learning tools classify communities into three groups roughly ordered by VC age and reach. Community VC financing is associated with faster maturation and greater innovation especially for early-stage firms without innovation histories.

Keywords: Venture capital, syndication, community detection, social networks, boundaries of the firm, machine learning

JEL Classification: G20, G24