Signaling and perceiving on equity crowdfunding decisions — a machine learning approach
利用五种机器学习模型，研究信号传递与投资者感知如何共同影响众筹成功，发现投资者更偏好结构化定量数据，且信息处理在降低信息不对称中起关键作用。
Abstract This study explores how signaling and perceiving jointly influence crowd investors’ decision-making. We utilize five machine learning models to assess the predictive power of various information types on crowdfunding success. Our findings indicate that investors prioritize well-structured quantitative data over complex qualitative content. Processing quantitative information is also found to be less cognitively taxing than extracting useful information from qualitative text and images. Entrepreneurs’ signaling and investors’ processing jointly reduce information asymmetry in crowdfunding, highlighting the critical yet often-overlooked role of investors’ information processing. Additionally, we test the policy effect of the ‘2016 Interim Measures on Online Lending ’ on crowdfunding success by comparing the predictive accuracy of information during the thriving and constraining periods of crowdfunding development in China. Our results have significant implications for policymakers that crowdfunding fosters economic growth by connecting entrepreneurs and investors and should not be halted due to risks, especially during periods of financial constraints.