Artificial Intelligence Adoption and the Within-Firm Labor Share: Evidence from Chinese Listed Firms
Using a staggered-adoption panel of Chinese listed firms and a novel text-based AI measure, this study finds that artificial intelligence adoption significantly reduces the within-firm labor share by 1.18 percentage points, driven by a decoupling of wages from productivity and rising markups, particularly in labor-intensive firms.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
For decades, economists watched a quiet but steady shift in how the money made by companies was divided between the people who work there and the people who own the business. In the past, this split remained surprisingly stable, but since the 1970s, the share going to workers has fallen in many places around the world. This change has sparked a major debate about what is driving it. Some experts believe the shift happens because the most successful companies are growing faster than everyone else, effectively changing the average without changing how any single company treats its workers. Others suspect that new technologies are fundamentally rewriting the rules inside individual firms, allowing them to keep a larger slice of the profits for themselves. The question that has remained difficult to answer is whether a specific technology, like artificial intelligence, is actually changing the internal balance of power within a single company, or if the story is more complex.
A new study by Jiawei Ma, a researcher at Macao Polytechnic University, looks directly inside the books of thousands of Chinese companies to find the answer. The research focuses on a specific period from 2010 to 2024, examining nearly 2,300 listed companies. To understand which of these companies were actually using artificial intelligence, the researcher did not rely on simple surveys or patent counts. Instead, a computer program was trained to read the detailed management sections of the companies' annual reports, searching for evidence that they had moved beyond just talking about technology to actually deploying it in their daily operations. This method allowed the researcher to pinpoint exactly when a company adopted these tools and to track what happened to its finances and workforce in the years that followed.
The findings reveal a clear and measurable change. When a company adopted artificial intelligence, the portion of its income that went to workers as wages dropped by about 1.18 percentage points. While this number might seem small, it represents a significant shift in the distribution of value created by the firm. The study shows that this decline is not because workers are being paid less or losing their jobs in a way that leaves them worse off. In fact, the opposite is happening for those who remain employed. After adopting artificial intelligence, companies saw their output per worker rise by nearly 6 percent, and the average wage for the remaining employees increased by about 2 percent. However, the wages did not keep pace with the surge in productivity. The workers were becoming more productive, but the extra value they created was not being passed on to them in the form of higher pay.
This gap between rising productivity and slower wage growth is the core of the discovery. The study explains that the decline in the labor share is driven by three main forces working together. First, the technology allows companies to replace certain tasks that humans used to do, reducing the total number of employees needed. Second, the boost in productivity allows the company to grow and become more efficient, but the financial benefits of this efficiency are largely retained by the company rather than shared with the workforce. Third, the companies that adopt these tools often gain more power in the marketplace, allowing them to raise their prices relative to their costs. This combination means that while the company becomes more profitable and the remaining workers earn slightly more, the overall share of the company's income going to labor shrinks.
The research also clarifies what is not happening. The data rules out the idea that workers are being impoverished by the technology; their wages are rising, just not as fast as the value they produce. It also challenges the notion that the decline in labor share is only a result of big companies growing while small ones shrink. The study proves that even within a single company, the act of adopting artificial intelligence changes the internal division of money. The effect is most pronounced in companies that rely heavily on human labor to begin with, where the shift away from workers is most visible. By using a clever comparison based on the history of telephone infrastructure in China to ensure the results are causal, the study provides strong evidence that artificial intelligence is a primary driver of this new economic reality. The result is a world where technology makes companies more productive and profitable, but the gains are distributed differently than in the past, with capital and profits capturing a larger share of the pie while labor's share recedes.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.