Data governance and green technological innovation performance: A curvilinear relationship.

The digital era has accelerated corporate innovation activity. Under the aim of "carbon neutrality", green technological innovation pushed enterprises toward resource conservation and reduced energy consumption. Although digital technology could increase the utilization of resources in the whole view, it remains unclear under what conditions the data affects green technological innovation performance (GTIP) effectively, and how data governance empowers different organizational management levels by delegating digital information in this progress. This study draws on the knowledge management theory and dynamic capability theory to explore the relationship between data governance and green technological innovation performance in firms. To examine our research model, annual reports of Chinese firms from 2013 to 2019 are analyzed. Results from Poisson fixed-effect model analysis indicate that data governance in general has no direct significant effect on green innovation performance, but for detail data governance items, green innovation reaches maximum, then decreases under the influence of decentralization of data governance, while a slightly U-shaped curve happens when centralization of data governance adopts, but these two curves are asymmetrical contradicted. Our analysis also highlights the important role of corporate environmental proactivity, as it positively moderates the relationship between two types of data governance and firm's green technological innovation performance. Besides, we examine this project by using three-dimensional plot method and Johnson-Neyman check, then find these regressions are robust in our analysis. Finally, implications and recommendations for data governance research and practices are presented. • Centralization of data governance shows slightly U-shaped relationship with GTIP. • Decentralization of data governance shows inverted U-shaped relationship with GTIP. • Corporate environmental proactivity moderates two non-linear relationships above. • Both granted patent and eliminated sample check support these results. [ABSTRACT FROM AUTHOR]