Computing Power and Labor Upgrading: Evidence from China's IDC License Panel Data (2015–2024)

This study utilizes a newly constructed panel that aligns China's administrative registry of Internet Data Center (IDC) operating licenses with firm-level financial, employment, and patent data from 2015 to 2024 to examine the relationship between computing-power development and labor-structure upgrading. A staggered difference-in-differences design considers the initial approval of an IDC license as a quasi-exogenous shock to firms' access to computing-power infrastructure, supported by event-study estimates that validate parallel pre-trends. The findings indicate that IDC license acquisition increases the proportion of highly skilled employees by approximately 2.6 percentage points and the proportion of highly educated employees by around 3.8 percentage points, reflecting increases of 11–12 percent relative to sample means. Mediation analysis reveals that increased innovation input and output, along with enhancements in total factor productivity and labor productivity, collectively represent approximately one-third of the overall effect. Further evidence indicates a decrease in the intensity of routine tasks and an increase in non-routine cognitive tasks, aligning with task-reallocation mechanisms. Heterogeneity analysis indicates that labor-upgrading effects are more pronounced among service-sector firms, coastal firms, and those situated in key-support intelligent-computing regions. The findings indicate that computing power serves as a digital general-purpose infrastructure, transforming internal production, innovation, and task allocation to favor high-skill and highly educated labor, thus intensifying the skill-biased nature of technological change in China's digital economy.