Importance of Recent Success and Initial Overconfidence: Factors of Crowdfunders' Limited Learning
by: Chankyung Pak and Rick Wash
Crowdfunding websites are new socio-technical systems where people raise money from large numbers of Internet users (“the crowd”) to support their projects. Crowdfunding shares many common elements with traditional funding mechanisms such as investment banks and public charities. Since many crowdfunders are non-professional in raising money and do not get helped by experts for their relatively small projects, they rely on learning from their experience to successfully raise money. In this paper, we examine the dynamics on a crowdfunding website DonorsChoose.org.
We present three behavioral traits of learning on DonorsChoose. First, success breeds more success, and failure breeds more failure. This trait is stronger than other types of learning such as information gaining or learning-by-doing on DonorsChoose. Second, the more recent the experience is, the more it affects the probability of success of the current project (recency effect). Lastly, as a result of discouraging effect of previous failure, initially over-ambitious teachers tend to drop out quickly rather than adjust their project size. Individual crowdfunders’ behavioral biases prevent them from becoming more efficient users of this innovative funding tool to some degree. On the other hand, it also suggests potential ways to improve their learning experience over time.
Chankyung Pak and Rick Wash. “Importance of Recent Success and Initial Overconfidence: Factors of Crowdfunders’ Limited Learning.” Poster in ACM Conference on Economics and Computation (EC). June 2014.