Abstract:
Whether data assets cause enterprises to fall into the new "productivity paradox" or ultimately enhance enterprise labor productivity remains a critical question. Based on the data of A-share listed companies on the Shanghai and Shenzhen Stock Exchanges from 2007 to 2023, this paper constructs a firm-level data asset index through the combined use of machine learning methods, text analysis, and sentiment polarity analysis to empirically examine the impact of data assets on labor productivity and the underlying mechanism. The results show a positive U-shaped relationship between data assets and labor productivity. The mediation analysis reveals that, in the initial stage of application, data assets may result in redundant human capital, declining innovation quality, and increased managerial power concentration, thereby reducing labor productivity. However, as the level of data asset application improves, data assets can enhance employees' overall competence, attract high-quality talents, foster a favorable environment for technological innovation and product iteration, and promote a flatter organizational structure with improved interdepartmental coordination and decision-making synergy, ultimately boosting labor productivity. The heterogeneity analysis shows that the positive U-shaped relationship between data assets and labor productivity is more pronounced in labor-intensive enterprises with moderate internal pay gap. This study provides evidence and insights for re-examining the "productivity paradox" in the context of the digital revolution and for guiding firms in building data assets.
下载:
