Herding in Equity Crowdfunding

Equity crowdfunding has recently become available and is quickly expanding. Concerns have been raised that investors ('backers') may be following the crowd 'too much' and making investments ('pledges') based on past investments rather than private information. We construct a model of equilibrium rational herding where uninformed investors follow signals generated by in formed investors with private information and a public belief generated by all past pledges. We show that large investments provide positive public information about the project's quality, whereas periods of absence of investment provide negative information. An information cascade is shown to occur only if not enough positive signals are generated. We then empirically analyse a large number of pledges from a leading European equity crowdfunding platform. We show that a pledge is strongly affected by both the size of the most recent pledge, and the time elapsed since the most recent pledge. For pledges that are not adjacent in the order of arrivals, the correlation between their sizes is still positive, but after being separated by two or more intervening pledges the correlation is no longer statistically significant. The effects are strongest for less-informed investors, and in some specifications the effects are strongest at the early stage of a campaign. We find similar results in IV analysis. Results are consistent with our model and inconsistent with some alternative models.