Assessing the impact of Sustainability disclosure in crowdfunding: a Natural Language Processing analysis.

This paper aims to initiate a dialogue between fintech and sustainability by evaluating the impact of sustainability disclosure on the outcomes of 370 Italian crowdfunding campaigns, collected from the Kickstarter platform between 2015 and 2024. Utilizing an NLP technique called zero-shot classification, we analyzed and classified the blurbs of the campaigns as sustainable, employing a rigorous decision framework. The model utilized was the BART-large model, trained with the MultiNLI dataset. The blurbs classified as sustainable by the model were further validated through human supervision to ensure accuracy. Furthermore, leveraging this classification, we constructed a measure of sustainability, and with a probit regression model, we assessed the impact of sustainability on the success of crowdfunding rounds. Several control variables were included such as the funding goal, campaign duration, blurb length, and the number of investors. The findings indicate that sustainability has a positive and significant impact. Disclosing sustainability information increases the probability of success by 7.73%. Additionally, the number of investors also has a positive and significant effect, while the funding goal has a negative effect. This paper contributes to the literature by providing evidence that sustainability acts as a signal of quality for investors, who are more inclined to invest in sustainable projects, resulting in a higher probability of success.