Automation, Employment Stability, and Labor Market Precarity among Low-Skilled Workers in China

• In Event: Inequality, Poverty, and Mobility Section Roundtables
  In Refereed Roundtable (60 minute): 1. Perspectives on Inequality in China

Abstract

Robots are not eliminating low-skilled workers from employment—they are keeping them in it under worse conditions. During China’s rapid automation expansion, employment rates among low-skilled workers remained stable. Yet stable employment—characterized by fixed employers, full-time hours, and formal contracts—declined sharply. The puzzle, then, is not why workers lost jobs, but why employment persisted while job security eroded.

This paper argues that automation reorganizes employment relations rather than simply reducing employment levels. Using three waves of the China Labor-force Dynamics Survey (CLDS: 2012, 2016, 2018), we compare automation’s effects on two outcomes: employment access (holding any full-time job) and employment stability (holding a secure, contracted position). Linking workers to sectoral robot adoption and accounting for global technological diffusion, we find a consistent contrast. Increases in robot density significantly reduce the probability of stable employment, with effects far larger for low-skilled workers and negligible for high-skilled workers. At the same time, automation modestly increases employment access, particularly among the low-skilled.

The same technological force erodes job security while sustaining labor market participation. Rather than displacing workers entirely, firms restructure employment relationships: as production becomes more capital-intensive, long-term workforce commitments give way to dispatch labor, short-term contracts, and informal arrangements. Automation thus redistributes risk within employment, weakening workers’ bargaining position while retaining them in the labor force.

These findings reframe the automation–inequality debate. The primary cost of technological change for low-skilled workers is not unemployment but job degradation—a quieter and less visible harm that aggregate employment statistics obscure. Understanding automation’s social consequences requires asking not only how many jobs exist, but what kinds of jobs remain.

Author

• Jiajia Wei, UMass Amherst