Information Frictions and Early-Stage Investors: Evidence from a Crowd-Rating Platform

Abstract

This paper seeks to investigate the existence and importance of collective intelligence to reduce information frictions by informing potential early-stage investors about venture quality. My context is an online platform where the community score projects, offering new insights into how the crowd affects subsequent venture success. I motivate my analysis using a statistical extraction model, which predicts that higher-scoring from the crowd signaling information about project quality, reducing information frictions to potential early-stage investors. To overcome the challenge of unobservables correlated with scoring, I leverage the quasi-random assignment of evaluators to project with different leniency, which leads to random variation in the overall score. Using this exogenous variation, I find no evidence that scoring from the crowd predicts subsequent venture success in the short and medium-run, but provides valuable information for entrepreneurs. In comparison, naive OLS estimates show positive correlations between the aggregate score and subsequent venture survival and employment, suggesting selection bias. Overall, my findings suggest that the crowd is unlikely to be an effective choice for revealing information about venture quality and reducing information frictions.