Abstract
This paper examines how informal and formal networks shape performance in the venture capital (VC) industry. Using data on all U.S.-based VC investments from 1990 to 2009 and career histories from LinkedIn, I develop a structural framework that connects three types of networks: coinvestment ties, historical affiliations, and latent social connections. In the baseline model, VC performance is a function of peer performance, capturing network spillovers through a micro-founded production function. To address endogeneity in network formation, I extend the model using a two-step instrumental variables strategy that leverages variation in past professional and alumni ties. Finally, I introduce an endogenous network formation model in which VCs strategically choose connections based on expected peer quality, allowing for the recovery of latent social networks from equilibrium outcomes. Across specifications, better-connected VCs exhibit significantly higher exit rates. Estimates from the endogenous model suggest that a 1% increase in social connectedness raises a VC's exit rate by 0.2 percentage points, while a 1% improvement in peer performance leads to a 0.74 percentage point increase in connection intensity. These findings highlight the economic value of informal relationships and offer new empirical tools for measuring network effects in private capital markets.