Does artificial intelligence bind labor market power? Evidence from listed manufacturing firms in China

Reducing firms' labor market power is crucial for improving labor-capital income distribution and mitigating labor market distortions. Based on data from Chinese listed manufacturing firms from 2011 to 2023, we estimate the impact of artificial intelligence (AI) on firms' monopsony power in the labor market. We first document that Chinese listed manufacturing firms possess significant labor market power, with an average markdown of 1.905, which implies that workers receive only 52.5% of their marginal output. AI penetration reduces firms' monopsony power in the labor market, although the magnitude is small, with an estimated elasticity of around − 0.01. This is because AI penetration reduces labor market power through human capital restructuring effects, while also increasing it through production enhancement effects. Moreover, this effect is more pronounced in firms with higher labor union expenditures, in state-owned enterprises, in industries having a higher market centralization or low-tech industries, and in regions with lower minimum wages or less developed factor markets. Further analysis reveals that AI raises workers’ average wages and reduces the dispersion of markdowns, thereby improving the efficiency of resource allocation.