ABSTRACT This study develops a novel city‐level metric for assessing digital government in China by applying an LSTM + RobotaBERT text classification model to local government procurement contracts. We investigate the impact of digital government on urban innovation, validating our metric against official digital development indices. Our analysis reveals a significant positive effect of digital government on the quantity of urban innovation, a finding that is robust to a comprehensive set of endogeneity checks. We then unpack the mechanisms, showing that this effect operates through two complementary channels: the direct reduction of firm‐level transaction costs and the broader improvement of the business environment. Crucially, our research demonstrates that the benefits of digital government are not unconditional; its positive impact on innovation is significantly stronger in cities with key enabling conditions, namely higher levels of human capital, financial development, and external cooperation. Furthermore, we find that digital government's influence extends beyond mere quantity to significantly enhance innovation quality, fostering more high‐cited, high‐knowledge‐width, and core technology patents. This work advances the understanding of digital government's role from a simple tool for efficiency to a powerful, context‐dependent catalyst for high‐quality urban innovation, providing nuanced insights for future policy.
Du et al. (Tue,) studied this question.