The concept of “high-quality development,” formally introduced during the 19th National Congress of the Communist Party of China in 2017, signifies a profound strategic shift in the nation’s economic paradigm. It marks a transition from a prior focus on the “quantity” of high-speed growth to a new era prioritizing the “quality” of economic expansion. Classical growth theory posits that economic advancement is fundamentally driven by factor inputs and technological progress. For a considerable period following its reform and opening-up, China’s growth was predominantly fueled by intensive factor accumulation, characterized by high investment and relatively low efficiency. However, against the backdrop of a diminishing demographic dividend and the unsustainability of perpetual high capital investment, achieving high-quality development necessitates a fundamental transformation of the growth engine—from reliance on factor inputs to dependence on technological progress. The primary agents of technological advancement are manufacturing enterprises within the real economy. Notably, technological progress in China’s manufacturing sector has historically exhibited a characteristic of being attached to capital investment. Therefore, embodied technological progress, where advanced technology is integrated into physical capital goods, becomes crucial. The application of industrial robots, which represent the fusion of material capital and cutting-edge technology, may serve as a pivotal catalyst for realizing high-quality development in China.
According to the International Federation of Robotics (IFR), an industrial robot is an automatically controlled, reprogrammable, multipurpose manipulator, programmable in three or more axes, which can be either fixed in place or mobile for use in industrial automation applications. It is designed to perform repetitive, complex, or hazardous tasks with high precision, potentially replacing human labor in such roles. Data reveals that China has emerged as the world’s largest market for industrial robots in terms of annual installations and operational stock. As a major developing economy undergoing a critical structural transformation, the proliferation of China robots is poised to exert a profound impact on productivity and, by extension, on the nation’s developmental quality.
This raises several critical questions: Can the adoption of robots genuinely promote the realization of China’s high-quality development objectives? Firstly, as a form of embodied technological progress, robot integration should increase the contribution share of technology to economic growth, aligning perfectly with the goals of high-quality development. Secondly, given the significant regional disparities in economic development levels and growth patterns across China, the impact of robot adoption on regional economic performance is likely to be heterogeneous. Finally, the effectiveness of embodied technology like robots is contingent upon a region’s absorptive capacity. Therefore, the pathway through which robot usage fosters high-quality development may involve enhancing and interacting with this regional capacity to digest and absorb new technology.
Literature Review and Theoretical Framework
Academic research on industrial robots primarily centers on two key themes: the impact on employment and the effect on productivity. The debate on employment is framed around the “displacement effect” versus the “reinstatement effect.” Studies, such as those by Acemoglu & Restrepo (2017), demonstrate that robots can substitute for human labor across a range of tasks, potentially depressing labor demand and wages. Conversely, other research argues that automation can create new tasks, industries, and job categories, leading to a net positive or polarized effect on employment. In the context of China, research has also explored the potential for robotics to mitigate challenges posed by an aging population.
On productivity, the consensus is markedly positive. Micro-level studies using firm data show that computerization and robot adoption significantly boost firm-level productivity, with effects amplifying over time. At the industry level, evidence confirms that robot density enhances labor productivity and value-added. Macroeconomic models incorporate robotics as a new form of capital that can influence long-run growth trajectories. However, there is a relative scarcity of research specifically investigating the impact of China robots adoption on regional “high-quality development,” particularly using a comprehensive efficiency metric. This study aims to fill this gap by examining how regional robot penetration influences technical efficiency, a core aspect of qualitative growth.
The theoretical foundation rests on endogenous growth theory and the concept of embodied technical change. We posit that robots are not merely capital but capital that embodies frontier automation and information technology. Their deployment directly alters the production function. The impact can be modeled by considering a region’s production frontier. Let a region’s aggregate output $Q$ be a function of conventional labor $L$, conventional capital $K$, and robot capital $R$. A simplified representation is:
$$ Q_{it} = A_{it} \cdot F(L_{it}, K_{it}, R_{it}) $$
where $A_{it}$ represents Total Factor Productivity (TFP) and $i$ and $t$ index region and time. Robot adoption affects $Q$ directly through $R$ and potentially indirectly by augmenting $A_{it}$ via spillovers and learning effects. High-quality development, emphasizing efficiency over mere scale, can be measured by how close a region operates to its production frontier given its inputs. This distance is captured by technical efficiency (TE).
Methodology and Measurement
Measuring High-Quality Development: Technical Efficiency
We employ Stochastic Frontier Analysis (SFA) to estimate regional technical efficiency, following the model of Battese & Coelli (1992). This method allows for the decomposition of output deviations from the frontier into random noise and technical inefficiency. We specify a Cobb-Douglas production function:
$$ \ln(Q_{it}) = \beta_0 + \beta_1 \ln(L_{it}) + \beta_2 \ln(K_{it}) + v_{it} – u_{it} $$
where:
- $Q_{it}$ is the real Gross Domestic Product (GDP) of province $i$ in year $t$.
- $L_{it}$ is the labor input, measured by the number of employed persons in urban units.
- $K_{it}$ is the capital stock, proxied by the accumulated fixed asset investment, using the perpetual inventory method with a depreciation rate.
- $v_{it}$ represents random statistical noise, assumed to be i.i.d. $N(0, \sigma_v^2)$.
- $u_{it} \geq 0$ represents technical inefficiency, a non-negative random variable.
The technical efficiency score for province $i$ at time $t$ is then defined as:
$$ TE_{it} = E[\exp(-u_{it}) | \epsilon_{it}] $$
where $\epsilon_{it} = v_{it} – u_{it}$. This score ranges between 0 and 1, with 1 indicating production on the frontier (fully efficient).
Measuring Robot Adoption: Robot Penetration
Following Acemoglu & Restrepo (2017), we construct a “robot penetration” index to measure the intensity of robot usage across Chinese provinces. Given the lack of direct provincial stock data from IFR for China, we utilize customs import data as a proxy. Industrial robots are identified under specific HS codes (e.g., 84795010, 84795090). The annual import value of these robots for each province is used to approximate the flow of new robot capital. A penetration rate normalizes this by the size of the workforce:
$$ Penetration_{it} (PR_{it}) = \frac{Robot\_Import\_Value_{it} (in 10,000 USD)}{Employment_{it} (in 10,000 persons)} $$
This gives a measure of robot capital per worker, indicating the intensity of adoption. Higher values signify greater integration of China robots into the provincial production system.
Econometric Model
To investigate the impact of robot penetration on high-quality development (proxied by technical efficiency), we specify the following panel data model:
$$ \ln(TE_{it}) = \alpha_0 + \alpha_1 \ln(PR_{it}) + \boldsymbol{\alpha_2 X_{it}} + \mu_i + \lambda_t + \epsilon_{it} $$
where:
- $\ln(TE_{it})$ is the natural logarithm of the technical efficiency score for province $i$ in year $t$.
- $\ln(PR_{it})$ is the core explanatory variable, the log of robot penetration.
- $\boldsymbol{X_{it}}$ is a vector of control variables including Foreign Direct Investment (FDI), Research and Development (R&D) expenditure, GDP per capita (to control for general development level), and the number of industrial enterprises.
- $\mu_i$ and $\lambda_t$ represent province and year fixed effects, controlling for unobserved time-invariant heterogeneity and common time shocks.
- $\epsilon_{it}$ is the idiosyncratic error term.
The coefficient $\alpha_1$ is our parameter of interest. A positive and significant $\alpha_1$ would indicate that higher robot penetration improves technical efficiency, supporting the hypothesis that China robots contribute to high-quality development.
Data and Descriptive Statistics
The analysis uses panel data from 29 Chinese provincial-level administrative units (excluding Tibet, Xinjiang, Hong Kong, Macao, and Taiwan) for the period 2017-2019. Data sources include provincial statistical yearbooks, China Customs statistics, and the National Bureau of Statistics. All monetary values are deflated to constant 2016 prices. The descriptive statistics for the key variables are presented below.
| Variable | Mean | Std. Dev. | Min | Max |
|---|---|---|---|---|
| Technical Efficiency (TE) | 0.75 | 0.12 | 0.52 | 1.00 |
| Robot Penetration (PR) (USD/Worker) | 9.85 | 23.45 | ~0 | 129.52 |
| Log(FDI) | 9.06 | 1.23 | 6.25 | 11.81 |
| Log(R&D Expenditure) | 4.63 | 1.07 | 2.34 | 7.06 |
| Log(GDP per Capita) | 3.91 | 0.20 | 3.34 | 4.30 |
| Log(Number of Enterprises) | 8.89 | 1.20 | 5.82 | 10.92 |
The calculated technical efficiency scores and robot penetration rates reveal significant regional disparities. The table below showcases the top and bottom performers.
| Province Category | High TE (Top 5) | TE Score | High PR (Top 5) | PR Score |
|---|---|---|---|---|
| High Performers | Liaoning | 0.996 | Shanghai | 99.81 |
| Shanxi | 0.874 | Jiangsu | 13.75 | |
| Hunan | 0.853 | Tianjin | 7.73 | |
| Hainan | 0.855 | Guangdong | 7.16 | |
| Ningxia | 0.820 | Heilongjiang | 6.09 | |
| Low Performers | Beijing | 0.523 | Yunnan | 0.0017 |
| Jilin | 0.630 | Guizhou | 0.0021 | |
| Qinghai | 0.635 | Hainan | 0.0066 | |
| Heilongjiang | 0.648 | Ningxia | 0.0049 | |
| Shaanxi | 0.653 | Henan | 0.0487 |
A clear stratification is observed in robot penetration. We define “High-Penetration Regions” (HPR) as the top 15 provinces and “Low-Penetration Regions” (LPR) as the remaining 14 based on a natural break in the distribution. This allows for heterogeneity analysis.
Empirical Findings and Analysis
Baseline Results
The baseline fixed-effects regression results confirm a significant positive relationship. After controlling for provincial and year effects as well as other factors, the coefficient on log robot penetration is positive and statistically significant. This suggests that a 1% increase in robot penetration is associated with an approximate 0.31% increase in technical efficiency, on average. This finding robustly supports the hypothesis that the adoption of China robots is a driver of high-quality development at the regional level.
| Variable | Full Sample | Low-Penetration Regions (LPR) | High-Penetration Regions (HPR) |
|---|---|---|---|
| ln(Robot Penetration) | 0.3079*** (0.057) |
0.1306*** (0.032) |
0.4941** (0.192) |
| Controls & Fixed Effects | Yes (FDI, R&D, GDPpc, Firms, Province FE, Year FE) | ||
| R-squared (Within) | 0.781 | 0.808 | 0.674 |
| Number of Observations | 78 | 42 | 45 |
| Note: *** p<0.01, ** p<0.05. Robust standard errors in parentheses. | |||
Heterogeneity Analysis: High vs. Low Penetration Regions
The stratified analysis reveals crucial heterogeneity. The impact coefficient is substantially larger for High-Penetration Regions (0.494) compared to Low-Penetration Regions (0.131). This difference is both economically and statistically significant. It indicates that the marginal benefit of an additional unit of robot penetration is much higher in regions that already have a strong foundation of robotics integration. This could be due to network effects, better complementary infrastructure (e.g., skilled maintenance, integrated supply chains), or stronger learning dynamics in HPRs. For LPRs, while the effect is positive, it is more muted, suggesting potential barriers to fully leveraging China robots, such as skill mismatches or inadequate complementary investments.
Addressing Endogeneity: Instrumental Variable Approach
To mitigate potential reverse causality (e.g., more efficient regions adopting more robots) and omitted variable bias, we employ an instrumental variable (IV) strategy. The instrument used is the annual number of academic publications related to “industrial robots” where the author’s affiliation is located in the province. The logic is that research activity is likely correlated with local industrial application and interest in robotics (relevance condition) but unlikely to directly affect aggregate provincial technical efficiency except through influencing robot adoption patterns (exclusion restriction). The two-stage least squares (2SLS) estimates, presented below, validate the baseline findings. The coefficient on robot penetration remains positive and significant, confirming a causal interpretation.
| Variable | Full Sample (2SLS) | First Stage F-Statistic |
|---|---|---|
| ln(Robot Penetration) | 0.5701** (0.269) |
25.7 |
| (All controls and fixed effects included) | ||
| Note: ** p<0.05. Strong instrument indicated by First Stage F-stat > 10. | ||
Mechanism: The Role of Absorptive Capacity
Technology diffusion theory emphasizes that the benefits of new technology are not automatic; they depend on the recipient’s absorptive capacity. We test this by introducing an interaction term between robot penetration and regional R&D intensity (a proxy for absorptive capacity). The extended model is:
$$ \ln(TE_{it}) = \gamma_0 + \gamma_1 \ln(PR_{it}) + \gamma_2 \ln(R\&D_{it}) + \gamma_3 [\ln(PR_{it}) \times \ln(R\&D_{it})] + \boldsymbol{\gamma_4 X_{it}} + \mu_i + \lambda_t + \epsilon_{it} $$
The results show that the interaction term $\gamma_3$ is positive and statistically significant across samples. This provides strong evidence that the positive effect of China robots on technical efficiency is magnified in regions with higher R&D intensity. The marginal effect of robot penetration is given by $\gamma_1 + \gamma_3 \ln(R\&D)$. In LPRs, where the standalone coefficient $\gamma_1$ is smaller, boosting R&D investment becomes a critical lever to amplify the gains from robot adoption. This implies that simply importing robots is insufficient; regions must invest in building complementary knowledge capital to fully absorb and exploit the technology.
Discussion and Implications for High-Quality Development
The empirical analysis demonstrates that the strategic deployment of industrial robots is a significant and robust contributor to enhancing regional technical efficiency in China, a core metric of high-quality development. The findings underscore several important policy implications and theoretical insights.
1. Robot Adoption as a Strategic Lever: The positive causal link validates the role of embodied technological advancement through robotics. Policymakers should view the promotion of the robotics industry—encompassing domestic manufacturing, strategic imports, and enterprise-level integration—as a strategic pillar for achieving quality-oriented growth. Fiscal incentives (e.g., accelerated depreciation for robotic equipment), tax benefits for R&D in automation, and targeted financial support can accelerate this transition.
2. Acknowledging and Addressing Regional Heterogeneity: The stark difference in impact between HPRs and LPRs highlights a “Matthew effect” in the dividends of automation. To prevent a widening of regional development gaps, a differentiated policy approach is essential:
- For High-Penetration Regions (HPRs): Policy should focus on moving up the value chain, supporting the development and adoption of next-generation, smart, and collaborative robots. Encouraging innovation in robotics software, AI integration, and system integration services will yield higher marginal returns and solidify their lead in high-quality development.
- For Low-Penetration Regions (LPRs): The priority is to overcome initial adoption barriers. This requires concerted efforts to reduce the effective cost of adoption for small and medium-sized enterprises (SMEs), perhaps through regional subsidy pools or leasing models. Crucially, our interaction analysis shows that building absorptive capacity is key. Therefore, policies for LPRs must jointly support robot acquisition and investments in relevant skills training, technical education, and applied research institutes focused on automation.
3. The Critical Nexus: Robots and Absorptive Capacity: The significant interaction effect between robot penetration and R&D intensity reveals the mechanism. High-quality development driven by China robots is not an automatic process. It is a function of: $$ \Delta TE \approx f( \underbrace{\text{Robot Capital}}_{PR}, \underbrace{\text{Knowledge Capital}}_{R\&D}, \text{Complementary Factors} ) $$ Regions that treat robot adoption as merely a capital investment project will see limited gains. Those that frame it as a technological capability-building exercise, coupling physical investment with sustained human capital and innovation system development, will capture the full benefits. This calls for integrated industrial and innovation policies.
4. Beyond Productivity: A Systems Perspective: While this study focuses on technical efficiency, the transition to a robot-intensive economy has broader implications consistent with high-quality development. It can improve product quality and consistency, enhance workplace safety by removing humans from dangerous tasks, and increase the flexibility of manufacturing systems. Moreover, by raising productivity, it contributes to higher wages and can help rebalance the economy towards consumption. However, managing the labor market transition through reskilling and social safety nets remains an imperative complementary policy to ensure inclusive growth.
Conclusion
This research provides empirical evidence that the adoption of industrial robots is a potent force for advancing China’s high-quality development agenda at the regional level. Using provincial panel data and robust econometric techniques, we find that robot penetration significantly improves technical efficiency, a key indicator of qualitative growth. However, this effect is highly heterogeneous, being markedly stronger in regions that already possess a high baseline level of robot integration. Furthermore, we identify a critical mechanism: the positive impact of robots is substantially amplified by a region’s absorptive capacity, proxied by its R&D intensity.
The policy implications are clear. A one-size-fits-all approach to promoting China robots is inadequate. A dual-track strategy is recommended: advancing frontier automation in leading regions while simultaneously fostering foundational adoption and capacity-building in lagging regions. For the latter, investments in robot technology must be bundled with investments in complementary knowledge capital—skills, training, and applied research. By doing so, China can harness the power of robotics not only to enhance productivity but also to promote more balanced, innovative, and sustainable regional development, thereby solidifying the path towards a high-quality economic future.