Unleashing the effect of data asset information disclosure on corporate investment efficiency: Fresh evidence from double-debiased machine learning

The current wave of the digital economy is predominantly fueled by cross-cutting technology areas, and data has become a fundamental resource and a critical factor of production for enterprises’ innovations and sustainable development. This paper examines the influence of data asset information disclosure on corporate investment efficiency (CIE) utilizing the double-debiased machine learning methodology. Employing a dataset of Chinese listed firms spanning 2010 to 2023, we uncover that information disclosure on data assets significantly enhances CIE. Specifically, the conditional average treatment effect of data asset information disclosure on corporate investment inefficiency is -0.314. The finding remains robust under various robustness tests. The causal mediation analysis reveals that this improvement occurs through enhanced information disclosure quality, reduced financing constraints, tax burden abatement, and fierce market competition. Furthermore, this enhancement effect is more significant in non-state-owned, non-mainboard-listed enterprises, and firms with highly readable annual reports. Further analysis shows that hard disclosures within data asset information have a stronger positive influence on CIE relative to soft disclosures. Moreover, the improvement in CIE further enhances firms' environmental, social, and governance (ESG) performance and boosts capital market efficiency. • We uncover the effect of data asset information disclosure on CIE via the DDML approach. • We measure data asset information disclosure via a text-mining approach. • Data asset information disclosure significantly enhances CIE. • The finding remains robust after various robustness checks. • We identify four potential channels within the nexus via causal mediation analysis.