Green oriented crowdfunding campaigns: Their characteristics and diffusion in different institutional settings

Highlights

• We study how country institutional setting affects the diffusion of green campaigns.

• We develop a machine-learning algorithm to identify “green” crowdfunding campaigns.

• Approximately 9.5% of all the campaigns presented on Kickstarter is “green”.

• “Green” campaigns significantly differ from the others along several dimensions.

• “Green” campaigns are common in countries with a lower sustainability orientation.

Abstract

In this paper, we investigate how the institutional setting of different countries affects the diffusion of green crowdfunding campaigns. To this aim, we develop and test two competing hypotheses about the association between country environmental sustainability orientation and the diffusion of green campaigns. To identify green campaigns, we develop an original machine-learning algorithm. We apply this algorithm to the population of 48,598 campaigns launched on Kickstarter between July 1, 2009 and July 1, 2012. Our findings show that green campaigns differ from others along several dimensions and are more diffused in countries with a limited environmental sustainability orientation. Implications for research, practice, and policy are discussed.

Introduction

An increasing body of research has shown that green entrepreneurial initiatives, defined as initiatives that have an environmental mission as their primal goal (Ortas et al., 2013; Thompson et al., 2011), face significant challenges in accessing financial resources (Fedele and Miniaci, 2010; Gaddy et al., 2017; Ghosh and Nanda, 2010; Lehner, 2013; Petkova et al., 2014; Ridley-Duff, 2009). In this scenario, crowdfunding potentially offers a new channel for green entrepreneurs to overcome these challenges. Previous studies have shown that differently from other entrepreneurial financial actors, backers of crowdfunding campaigns are not moved by profit motivations alone and select projects based also on their wish to help others and to support causes they care for (Belleflamme et al., 2014; Gerber and Hui, 2013). Moving from these arguments, a number of recent studies has investigated green initiatives launched on crowdfunding platforms, focusing on the likelihood of success of crowdfunding campaigns with an environmental focus (e.g., Bartenberger and Leitner, 2013; Bonzanini et al., 2015; Calic and Mosakowski, 2016; Candelise, 2016; Hörisch, 2015; Lam and Law, 2016). Comparatively, little attention has been given to factors that influence the propensity to launch green initiatives on crowdfunding platforms. This is an important gap in the literature given the crucial role that crowdfunding may play in financing green entrepreneurial initiatives. The aim of this study is to contribute to filling this gap.

For this purpose, we take inspiration from the view that the institutional setting contributes to inform the behavior of individuals and notably of entrepreneurs (Bruton et al., 2010). Particularly, we consider the environmental sustainability orientation of the institutional setting in a given country. We define the environmental sustainability orientation as the presence of formal (i.e., policy and regulation) and informal (i.e., norms, values, beliefs and practices) institutions,¹ that are concerned not only with the current level of economic and non-economic well-being but also with its sustainability over time i.e. the ability to pass natural, physical, human, and social resources to future generations (Stiglitz et al., 2010). We argue that the environmental sustainability orientation of the institutional setting influences both entrepreneurs' willingness to launch a green-oriented entrepreneurial initiative and their inclination to use crowdfunding to finance such initiatives. In countries where the institutional setting has a stronger environmental sustainability orientation, one may expect that there are more green-related business opportunities. Moreover, as green initiatives are more legitimized, entrepreneurs may be more inclined to resort to crowdfunding to finance their green initiatives as they anticipate a greater likelihood of success of their campaigns. However, in these countries, entrepreneurs may also find it easier to finance their green initiatives through traditional channels. The ease of access to traditional channels may have a direct negative effect on their inclination to use crowdfunding. Moreover, because of adverse selection, green crowdfunding campaigns may be perceived by potential backers as having low quality, which would make entrepreneurs even more reluctant to finance their green initiatives through crowdfunding. Considering that opposed forces are at work, it is difficult to argue a priori whether the environmental sustainability orientation of a country's institutions is positively or negatively associated with the diffusion in this country of green-oriented crowdfunding campaigns. In this paper, we investigate the relative explanatory power of these competing hypotheses.

Answering this research question raises a serious methodological problem: identifying green-oriented campaigns. Prior studies have accomplished this task by either focusing on a limited number of campaigns and providing a subjective evaluation of their green orientation (Calic and Mosakowski, 2016) or using the classification of green campaigns provided on crowdfunding platforms (e.g., Hörisch, 2015). These strategies have important drawbacks. The subjective evaluation of green campaigns based on the textual description of the project has the advantage of being very precise (Kononenko and Bratko, 1991), particularly if made by a sufficient number of well-trained researchers, as in the case of Calic and Mosakowski (2016). However, this approach is only applicable to a limited sample of campaigns, due to the considerable time required to perform the evaluation. At the same time, using the classification of green campaigns provided by platforms has the clear advantage of relying on an easy to access and publicly available classification. However, it entails the risk of overlooking a large number of green-oriented campaigns. Consider, for instance, a movie produced through environmental sustainable techniques or a piece of art created with the intent of warning people about an environmental threat. It is hard to argue that these initiatives are not green. This notwithstanding, they are unlikely to be labeled as green by platforms, considering that platforms typically provide a single label to a campaign (see e.g., Mollick, 2014). An interesting step forward is the approach by Cumming et al. (2017). They implement a computer-based text analysis technique aimed at searching for a predefined set of related keywords in the description of the focal project. The accuracy of the classification clearly depends on the comprehensiveness and uniqueness of the chosen set of keywords. In this paper, we considerably improve this methodology by employing a robust machine-learning algorithm to create a content-specific classifier of green-oriented crowdfunding campaigns, based on the project descriptions posted on the crowdfunding platform. Compared to the keywords-based algorithm implemented in Cumming et al. (2017), the machine-learning approach offers a great advantage. Indeed, the set of most informative keywords used to discriminate between green and non-green campaigns is automatically extracted from a large training corpus of labeled descriptions of crowdfunding projects. Therefore, keywords are not selected in advance, but they are learnt from the specific textual domain.

We apply the classifier to the population of 48,598 crowdfunding campaigns launched on Kickstarter in the period between July 1, 2009 and July 1, 2012. Approximately 9.5% of all the campaigns presented on Kickstarter are identified by the algorithm as green initiatives. Our findings show that these campaigns significantly differ from other campaigns along several dimensions. Specifically, green crowdfunding campaigns have a larger capital goal, provide more information (both visual and verbal), are launched by creators with significantly larger amounts of social capital developed within the crowdfunding platform (Colombo et al., 2015), and a smaller external network (Mollick, 2014). More importantly, our econometric estimates show that green campaigns are proportionally more common in countries presenting a lower environmental sustainability orientation.

The paper is organized as follows. Section 2 discusses the theoretical background and develops hypotheses. In Section 3, we present data, methodology and descriptive statistics. Next, we present the main Results. The last section concludes the paper and discusses the implications for practice, policy, and future research.