The application of intelligent robots is profoundly transforming human production and lifestyles, with their penetration into traditional labor forces exhibiting dual effects of substitution and creation. This dynamic dual effect not only drives productivity enhancement and occupational transformation through intelligent development, but also introduces greater uncertainty to employment patterns and income distribution in the labor market, manifesting as the “hollowing out of the middle class” and emerging income inequality issues. On one hand, artificial intelligence has led to the replacement of numerous simple, repetitive tasks by robots, intensifying unemployment risks for medium- and low-skilled workers or forcing them into lower-paying jobs. On the other hand, intelligent penetration fosters the emergence and development of new professions demanding higher skills and knowledge. These transformations pose new challenges to labor market stability and equity. Therefore, examining the micro-level effects of robot penetration and ensuring that artificial intelligence development benefits broader social groups holds significant practical importance.
This study constructs a task-biased theoretical model integrating horizontal penetration and vertical innovation to analyze impacts of intelligent robots on labor income, wage rates, and labor participation rates at the micro level. The theoretical analysis suggests that current intelligent robot penetration reduces labor income and displaces labor participation share while increasing wage rates, thereby widening income disparities between high- and low-skilled sectors. Empirically, utilizing data from the International Federation of Robotics (IFR) and the China Household Finance Survey (CHFS), we constructed a micro-level dataset comprising 7,674 observations from 2015-2019. Employing a two-way fixed effects model, we systematically examined the actual impacts and mechanisms of intelligent robot penetration on labor income, wage rates, and labor participation share (measured by working months and days), addressing endogeneity through instrumental variable approaches. Heterogeneity analyses based on occupational and industrial variations were conducted. The findings demonstrate that robot penetration leads to decreased labor income and labor participation share alongside increased wage rates. Heterogeneity analysis reveals that such penetration exacerbates the “polarization effect” between high- and low-tech sectors while narrowing participation rate disparities. Mechanistic analysis indicates that robot penetration affects labor employment through occupational mobility, industrial transformation, and increased employment stability risks. Furthermore, the analysis shows that increased penetration rates widen income inequality, particularly in service sectors.
Compared to previous studies, the contributions of this paper are primarily reflected in: (1) Examining the dynamic evolution of intelligent robots’ substitution and creation effects, particularly the structural differentiation characteristics among occupational groups at the micro level. (2) Integrating income and labor participation within a unified framework, focusing on income distribution and labor force participation to accurately assess the comprehensive economic impact of intelligent robot penetration, thereby expanding the dimensionality of impact evaluation at the micro level. (3) Systematically discussing the synergistic mechanisms through which intelligent robot penetration generates both positive and negative impacts on the labor market, thus providing more precise decision-making support for labor market policy formulation in the artificial intelligence era.
This study systematically validates the comprehensive effects of intelligent robot penetration on the labor market, offering a micro-level perspective for understanding emerging challenges in the artificial intelligence era. Accordingly, policy recommendations are proposed to promote the efficient synergy and integrated advancement of artificial intelligence technology development and high-quality full employment. These include achieving the transition from “income polarization” to “opportunity equity” through workforce skills training and structural fiscal and tax policies, and shifting from “passive safety nets” to “proactive empowerment” to smoothly navigate the coexistence of upward labor mobility and increased employment risks.
PRETTNERK, STRULIKH. Innovation, automation, and inequality: policy challenges in the race against the machine[J].Journal of Monetary Economics,2020,116:249-265.
[2]
ACEMOGLUD, RESTREPOP. Robots and jobs: evidence from US labor markets[J].Journal of Political Economy,2020,128(6):2188-2244.
[3]
ACEMOGLUD, RESTREPOP. Automation and new tasks: how technology displaces and reinstates labor[J].Journal of Economic Perspectives,2019,33(2):3-30.
[4]
JONESB F, LIUX J. A framework for economic growth with capital-embodied technical change[J].American Economic Review,2024,114(5):1448-1487.
LINC, CHENX L, CHENW Z, et al. Artificial intelligence,economic growth and household consumption improvement: a capital structure perspective[J].China Industrial Economics,2020,37(2):61-83.
KONGG W, LIUS S, KONGD M. Robots and labor employment:an empirical investigation based on heterogeneity of industries and regions[J]. China Industrial Economics,2020,37(8):80-98.
[13]
ACEMOGLUD, RESTREPOP. The race between man and machine: implications of technology for growth, factor shares, and employment[J]. American Economic Review,2018,108(6):1488-1542.
GUOK M, WANGY B, GONGL T. Labor supply transformation, well-functioning government and the era of artificial intelligence[J].Journal of Management World,2023,39(6):1-21.
WANGY Q, DONGW. How the rise of robots has affected China’s labor market:evidence from China’s listed manufacturing firms[J].Economic Research Journal,2020,55(10):159-175.
[18]
GRAETZG, MICHAELSG. Robots at work[J].The Review of Economics and Statistics,2018,100(5):753-768.
LIUY, HANY H, WANGX B. Can industrial intelligentization balance the promoting growth and ensuring living needs?[J].Journal of Quantitative & Technological Economics,2023,40(6):69-90.
HUANGH Q, DAIT S, SHENJ. AI, learning-by-doing and skill premium: a study based on a model of endogenous technological change[J].China Industrial Economics,2024,41(2):99-117.
SUNZ, GAOX K. Intelligentization, labor effective allocation, and the improvement of employment quality[J].Journal of Xi’an Jiaotong University(Social Sciences),2025(1):53-68.
WANGL H, QIANY Y, SONGD L, et al. The job transition effects and employment sensitivity among specific groups of robot applications: empirical evidence from a micro-individual level perspective[J].Economic Research Journal,2023,58(7):69-85.
[38]
BLUNDELLR, CHIAPPORIP A, MAGNACT, et al. Collective labour supply: heterogeneity and non-participation[J].Review of Economic Studies,2007,74(2):417-445.
[39]
BLANCHARDO, FISCHERS. Lectures on macroeconomics[M].Cambridge: MIT Press,1989:337-542.
ZHAOT, ZHANGZ, LIANGS K. Digital economy, entrepreneurship, and high-quality economic development:empirical evidence from urban China[J].Journal of Management World,36(10):65-76.
[42]
COHNJ B, LIUZ, WARDLAWM I. Count (and count-like) data in finance[J].Journal of Financial Economics,2022,146(2):529-551.
[43]
BARTIKT J. Who benefits from state and local economic development policies?[M]. Kalamazoo: W.E. Upjohn Institute for Employment Research,1991:81-286.
[44]
BAETSCHMANNG, BALLANTYNEA, STAUBK E, et al. Feologit: a new command for fitting fixed-effects ordered logit models[J].The Stata Journal,2020,20(2):253-275.
GUOH, FENGX Y. Multiple digital divides and household income disparities: theoretical framework and empirical evidence[J]. Economic Review Journal,2024,40(2):79-90.
LIUW J, HAOY Z. Digital divide and labor reallocation: analysis of dynamic effects and efficiency improvement[J]. Modern Economic Science,2025,47(3):18-34.