In recent decades, China has emerged as a global export powerhouse, with its share of world trade rising significantly. However, challenges such as demographic shifts, rising production costs, and external economic pressures have slowed export growth, prompting a search for new competitive advantages. The application of industrial robots, often referred to as “China robots,” is seen as a key driver in this context, potentially replacing traditional labor advantages with automation红利. This article explores how the adoption of China robots influences export scale and structure, focusing on the role of factor mobility. We analyze data from 2006 to 2016, employing empirical models to assess these effects and discuss policy implications for sustainable trade development.
Theoretical analysis suggests that China robots can enhance export performance through several channels. First, automation improves productivity, enabling firms to overcome export fixed costs and expand overseas sales. This leads to our first hypothesis: H1: Industrial robot application promotes export scale expansion, particularly in cities with higher robot density. Second, China robots may alter factor flows—attracting high-skilled labor and capital while displacing low-skilled workers, thereby upgrading local human capital and fostering manufacturing agglomeration. We propose: H2: Robot application boosts exports by attracting high-skilled labor and improving human capital. H3: Robot application drives exports by facilitating capital transfer and industrial clustering. These mechanisms underscore the transformative potential of China robots in reshaping trade patterns.
To test these hypotheses, we construct a city-level dataset combining robot statistics from the International Federation of Robotics (IFR) with Chinese customs export records. Our core explanatory variable, robot density, is measured as robots per 10,000 workers, calculated using pre-sample industrial employment structures and industry-level robot installations. The model specification is as follows:
$$ \ln(\text{exp}_{ijt}) = \beta_0 + \beta_1 \ln(\text{robots}_{it}) + \sum_k \beta_k X_{kit} + \mu_i + \upsilon_j + \varphi_t + \varepsilon_{ijt} $$
Here, \( \text{exp}_{ijt} \) denotes the export value of product \( j \) in city \( i \) at year \( t \), \( \text{robots}_{it} \) is the robot density, and \( X \) includes control variables like GDP per capita, labor costs, informatization, financial development, and FDI. We include city, product, and year fixed effects to account for unobserved heterogeneity. Robust standard errors are clustered at the city level.
Table 1 summarizes the baseline regression results for export scale and margins. The findings indicate that China robots significantly boost city-level export value, with the effect concentrated in the quantity margin rather than price or extensive margins.
| Dependent Variable | Coefficient on Robot Density | Controls | Fixed Effects | Observations | R² |
|---|---|---|---|---|---|
| Export Value (ln) | 0.174** (0.073) | Yes | Yes | 2,815,468 | 0.340 |
| Export Quantity (ln) | 0.202** (0.080) | Yes | Yes | 2,815,468 | 0.547 |
| Export Price (ln) | 0.031 (0.047) | Yes | Yes | 2,815,468 | 0.791 |
| Export Destination Count (ln) | 0.041 (0.066) | Yes | Yes | 2,815,468 | 0.462 |
To address endogeneity concerns, we employ an instrumental variable approach using U.S. industry-level robot installations as an instrument for China robots. The results, shown in Table 2, confirm a positive and significant impact, supporting hypothesis H1. Robustness checks, such as using robot stock data or excluding the automotive sector, yield consistent findings, highlighting the reliability of the China robots effect.
| Stage | Variable | Coefficient | Standard Error |
|---|---|---|---|
| First Stage | U.S. Robot Density | 0.285*** | 0.089 |
| Controls | Yes | – | |
| Second Stage | Robot Density | 1.303* | 0.684 |
| Kleibergen-Paap F Statistic | 19.808 | – |
Heterogeneity analysis reveals nuanced effects of China robots on export structure. For trade modes, robots promote both processing and ordinary trade, but with a stronger impact on processing trade, potentially slowing the transition to higher-value activities. Regarding product technology, robots significantly increase exports of high-tech manufactured goods, as shown in Table 3, indicating an optimization of export composition.
| Product Type | Coefficient on Robot Density | Observations | R² |
|---|---|---|---|
| Primary Products | -0.016 (0.087) | 144,210 | 0.290 |
| Resource-based Manufactures | 0.112 (0.077) | 433,144 | 0.275 |
| Low-tech Manufactures | 0.180** (0.075) | 1,128,387 | 0.399 |
| Medium-tech Manufactures | 0.129** (0.056) | 837,291 | 0.343 |
| High-tech Manufactures | 0.304*** (0.086) | 249,495 | 0.413 |
Export market analysis shows that China robots boost sales to both OECD and non-OECD countries, with a slightly larger effect on non-OECD markets, but no significant change in market share distribution. This suggests that while robots expand reach, they do not alter market structure substantially.
Mechanism tests explore factor mobility. For labor, we estimate a conditional logit model for migrant workers. The results in Table 4 indicate that China robots attract high-skilled labor while potentially displacing low-skilled workers, thereby enhancing local human capital and supporting H2.
| Labor Type | Coefficient on Robot Density | Pseudo R² | Observations |
|---|---|---|---|
| High-skilled Migrants | 0.163*** (0.017) | 0.079 | 31,800 |
| Low-skilled Migrants | -0.113 (0.151) | 0.067 | 162,339 |
For capital, we examine new firm entry and industrial share. Table 5 shows that robot application increases the number of new industrial enterprises and elevates manufacturing concentration, confirming H3. This capital transfer fosters agglomeration economies, further driving exports.
| Capital Measure | Coefficient on Robot Density | Observations | R² |
|---|---|---|---|
| New Enterprise Count | 0.102*** (0.030) | 2,970 | 0.242 |
| Manufacturing Share (Firms) | 0.015*** (0.005) | 2,970 | 0.873 |
| Industrial Output Share | 0.008** (0.003) | 2,970 | 0.761 |
Spatial externalities are assessed using a Spatial Durbin Model (SDM) with geographic adjacency and inverse distance matrices. The direct and indirect effects decomposition in Table 6 reveals that while China robots boost local exports, they suppress neighboring city exports due to factor siphoning. However, the net total effect remains positive, underscoring the overall benefits of robot adoption.
| Spatial Effect | Adjacency Matrix | Inverse Distance Matrix |
|---|---|---|
| Direct Effect | 0.158*** (0.051) | 0.177*** (0.050) |
| Indirect Effect | -0.096*** (0.027) | -0.064*** (0.018) |
| Total Effect | 0.062*** (0.026) | 0.113* (0.060) |
In conclusion, the widespread use of China robots has significantly propelled export growth, primarily through quantity expansion and factor mobility channels. These robots attract high-skilled labor and capital, leading to industrial clustering and enhanced productivity. However, regional disparities may widen due to siphoning effects, and trade structure adjustments, such as a shift from processing to ordinary trade, require further stimulation. Policymakers should promote robot diffusion with regional coordination, support skill development, and incentivize high-value exports to harness the full potential of China robots. This transition from demographic to robotic dividends offers a viable path for sustaining China’s export competitiveness in the global arena.