Research on Digital Transformation Alleviating Financing Constraints and Promoting Green Innovation Through Semantic Measurement and Causal Forests of Large Language Models

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Digital transformation transforms an enterprise's green innovation capabilities, requiring the measurement and identification of genuine organizational changes and analyzing how capital constraints shape the marginal returns of technological investment. This paper, based on 18,642 annual observations of Chinese A-share listed companies, uses an open-source Chinese large language model to construct a digital transformation index. With financing constraints as the transmission variable, and using the five-fold cross-fitting causal forest to identify the conditional average treatment effect. The results show that digital transformation significantly enhances green innovation and reduces financing constraints; the Bootstrap indirect effect is 0.044, accounting for 23.66% of the total effect. Financing constraints are the most important condition variable in heterogeneous effects, and the digital transformation effect increases from the low constraint group to the higher constraint group, with a slight decline at the highest quantile but no significant difference. The study can reveal the resource boundaries masked by the average effect while maintaining economic implications.