信息摩擦、公共数据与企业创新边界

Abstract

In the digital era, public data has become a vital infrastructure for firm innovation. While governments globally are expanding public data sharing, how such data influences firms' strategic innovation, particularly under information frictions, remains poorly understood. In China, public data openness primarily involves low-frequency, unstructured information such as policy documents or industrial plans. Although these data are often unsuitable for direct algorithmic or production use, they provide substantial informational value by clarifying policy priorities, reducing uncertainty in the innovation landscape, and helping firms anticipate directional shifts in technological and regulatory environments. This paper investigates how public data influences the expansion of firms' innovation boundaries from the perspective of information frictions. Theoretically, we develop a model in which firms face uncertainty in R&D investment and must choose between incremental and radical innovation strategies. The model introduces a Bayesian signal environment where firms form expectations about the innovation landscape using both private and public data. Public data primarily operates through two channels: first, an information channel, whereby its informational content enhances firms' ability to anticipate innovation fundamentals and reduces uncertainty in R&D investment; and second, a coordination channel, whereby its shared nature allows multiple firms to utilize it collectively, mitigating inter-firm information asymmetries and improving collaborative efficiency. Empirically, we exploit the staggered rollout of municipal public data platforms in China as a quasi-natural experiment. The results show that the launch of public data platforms significantly promotes firms' radical innovation. Mechanism analyses further reveal that public data access reduces R&D investment volatility and guides firms' technological deployment toward policy-relevant domains, supporting the information channel. In addition, public data facilitates inter-organizational collaboration by enhancing coordination, as evidenced by increased joint patenting, thereby validating the coordination channel. This study makes three main contributions. First, it develops a theoretical model of heterogeneous innovation choices under information frictions, integrating both private and public data into firms' R&D decision-making. Second, this paper offers a new perspective on the determinants of radical innovation. Existing studies have focused on factors such as managerial characteristics, government subsidies, and technological spillovers. In contrast, this study emphasizes the information environment as a key driver of radical innovation by identifying the information and coordination effects of public data. Third, it contributes to the growing literature on public data policy evaluation. This paper employs large language models to extract and classify unstructured content from both patent texts and public data directories. This enables a novel measure of topic-level alignment between government-supplied data and firms' innovation activities, offering new evidence on the micro-level channels through which public data affects firm behavior. This paper advances the interdisciplinary research at the intersection of information economics and innovation boundary theory, providing a solid theoretical foundation and empirical evidence for understanding how data, by alleviating information frictions, reconfigures firms' micro-level innovation decisions. Furthermore, the findings offer targeted policy insights for fully unlocking the informational value of public data and optimizing the functional design of government data openness platforms. The study also offers novel perspectives and practical implications for leveraging data as a key factor to empower high-quality economic development, modernize national governance, and advance digital government development.