Chen Ouyang
Sociological Science August 31, 2026
10.15195/v13.a41
Abstract
In the United States, overwork migrated up the class structure, and its hourly wage association turned from penalty to premium in aggregate estimates. The “996” controversy suggests China is replaying this elitization. I test both with 11 waves of the Chinese General Social Survey (2006–2023; 36,282 nonfarm workers), using wave-specific, subperiod, and full-period estimates. Neither is supported as a broad, durable pattern. Overwork stays concentrated among manual, service, and self-employed workers; a post-2015 rebound was not disproportionately professional—manual workers rose at least as much—and was not detectably sustained in 2023. Overwork does not pay: overworkers log two-thirds more hours than comparable workers yet earn no more annually, and the hourly penalty—deepest for blue-collar employees—narrowed partially after 2010 and remained negative in every wave. State-sector employment is associated with lower overwork throughout. Elitized overwork appears institution-dependent; the sample instead shows a non-elite overwork configuration.
In the United States, overwork migrated up the class structure, and its hourly wage association turned from penalty to premium in aggregate estimates. The “996” controversy suggests China is replaying this elitization. I test both with 11 waves of the Chinese General Social Survey (2006–2023; 36,282 nonfarm workers), using wave-specific, subperiod, and full-period estimates. Neither is supported as a broad, durable pattern. Overwork stays concentrated among manual, service, and self-employed workers; a post-2015 rebound was not disproportionately professional—manual workers rose at least as much—and was not detectably sustained in 2023. Overwork does not pay: overworkers log two-thirds more hours than comparable workers yet earn no more annually, and the hourly penalty—deepest for blue-collar employees—narrowed partially after 2010 and remained negative in every wave. State-sector employment is associated with lower overwork throughout. Elitized overwork appears institution-dependent; the sample instead shows a non-elite overwork configuration.
 | This work is licensed under a Creative Commons Attribution 4.0 International License. |
Supplemental Materials
Reproducibility Package: All analysis code, the cross-wave harmonization pipeline, archived numeric outputs, and documentation needed to reproduce every empirical result in this article are deposited at Zenodo: https://doi.org/10.5281/zenodo.21709718. The CGSS microdata cannot be redistributed under the CNSDA data-use agreement; the package documents, wave by wave, how registered researchers can obtain each file free of charge from the Chinese National Survey Data Archive (cnsda.org).
- Citation: Ouyang, Chen. 2026. “Overwork Without Premium: The Persistent Class Gradient of Long Hours in China, 2006–2023” Sociological Science 13: 1079-1106.
- Received: July 12, 2026
- Accepted: August 8, 2026
- Editors: Ari Adut, Wei-hsin Yu
- DOI: 10.15195/v13.a41