Catching Outliers: Committee Voting and the Limits of Consensus When Financing Innovation

We provide novel evidence that a substantial share of venture capital investors in the U.S. use a voting model where a single partner 'championing' an early stage investment is sufficient for an investment committee to do the deal, even if other partners are not as enthusiastic. Their stated reason for this voting rule is to 'catch outliers'. The same VCs also tend towards more conventional 'majority' or 'unaninimity' rules for later stage investments. We analyze this evidence through the lens of several conventional models of information aggregation in committees, and conclude that it points to a model in which different voting partners get signals about different aspects of the project and superstar projects are the ones that excel on some dimensions even if potentially awed on others. In this case, if the distribution of investment returns is sufficiently heavy-tailed, a champions rule is optimal, while for less heavy-tailed distributions, more consensus is optimal. We then show empirically that the distribution of early stage returns have significantly heavier tails than late stage returns, validating the model. In a quantitative example, we find that a majority voting rule in early stage investing reduces the chances of finding the best investments by up to 70% relative to a 'champions' model.