Rivaling reinforcement and resource reallocation: How do industrial robots enhance company’s innovation?

Main results

To explore how IBs influence company innovation capabilities, this study employs data from industrial firms listed on the A-share market spanning 2011 to 2019. Estimations are conducted using both OLS and IV methods, with results presented in Table 3. Columns (1) and (3) employ the number of patent applications to measure innovation, while columns (2) and (4) use the number of granted patents. Due to the zero values issue in patent applications, this study follows common practice in the literature by adding 1 to the patent count and taking the logarithm in columns (1) and (2), and applying an arcsine hyperbolic transformation in columns (3) and (4). Additionally, to control for time trends and company-specific characteristics, fixed effects for both companies and time are included in columns (1) to (4). Panel A utilizes the OLS estimation method. Results in columns (1) to (4) all show positive coefficients for IBs, which are statistically significant, indicating that the use of IBs significantly enhances company innovation capabilities, thus confirming hypothesis H1.

However, company innovation might reverse impact the use of IBs. Companies with strong innovation capabilities may be more inclined to adopt IBs. Innovation drives automation, increased productivity, and reduced labor costs. The demand for more accurate and efficient production processes, as well as improvements in production, management, or operational processes, increases with innovative activities. We adopt the IV estimation method to reduce the endogenous bias caused by reverse causality (Table 3, Panels B and C). Panel B presents the results of the reduced form, demonstrating the direct impact of IB penetration in relevant U.S. industries on the innovation of Chinese industrial companies. The results indicate that IB penetration in related U.S. industries significantly enhances the innovation of Chinese companies, reflecting the direct impact of exogenous industry technology shocks on the innovation of China’s listed industrial companies. Furthermore, the reduced form estimates are similar to the OLS results, suggesting that corporate innovation has little reverse impact on the use of IBs. The coefficients of control variables are shown in Appendix Tables A1 and A2.

Panel C displays the results of the 2SLS estimates. Column (5) of Panel C lists the first-stage estimation results. The estimated coefficient of the IV on the use of IBs in Chinese industrial companies is 0.784, which is significant at the 1% level. This suggests a penetration effect of U.S. IBs on Chinese company’s IB adoption. Columns (1) to (4) of Panel C report the second-stage estimates. The results show that IBs have a significant promotional effect on company innovation. A 1% increase in IB penetration in Chinese industrial companies significantly boosts patent grants and filings by about seven, which is consistent with the conclusion in Panel A. Additionally, the Kleibergen-Paap (KP) Wald rk F-statistic of 877.755 exceeds the Stock-Yogo critical value of 16.38. The adjusted R² in the first stage is 0.953, indicating that U.S. IB penetration explains most of the changes in IB penetration across industries in China, demonstrating strong correlation and not being a weak instrument variable.

Combining similar estimation results from the reduced form and OLS, this study concludes that China is keeping up with the technological advancements of U.S. industries in terms of IB utilization. It is worth noting that the estimated coefficients in Panel C are slightly larger than those in Panel A. This discrepancy may be due to Panel A excluding most of the endogeneity through fixed effects and control variables, but still being susceptible to unobservable factors such as time-varying industry shocks and entrepreneurial factors. In contrast, using IB penetration in relevant U.S. industries as an IV provides external industry technology shocks, helping to identify the causal impact of exogenous IBs shocks on companies.

Furthermore, we examine the heterogeneous impact of IBs on innovation among companies of varying ownership types. This analysis is underpinned by a thorough comprehension of the ownership structure of Chinese companies, recognizing that SOEs may have greater access to government resources and support, while private companies may exhibit greater flexibility and proactivity in pursuing efficiency and innovation. By elucidating these differences, our study highlights the ownership-specific factors that policymakers need to consider and suggests how policies can be designed to promote technological innovation more equitably (Appendix D, Table D1–5).

Robustness test

To further ensure the robustness of our findings, we consider the possibility that the impact of industrial robot adoption on firm innovation may not be immediate. Technological adoption and its influence on organizational processes typically require time to materialize. Insight of this, we conduct an additional robustness test by using the one-period lag of the IB adoption. The estimation results, presented in Table 4, show that the coefficients on the lagged IB adoption remain positive and statistically significant. Moreover, the magnitude of the effects is consistent with the baseline results, reinforcing the reliability of main conclusions.

Mechanism analysis

Market competitiveness of companies

Firms’ competitiveness is primarily reflected in cost control, profitability, and productivity. Cost control is the company’s ability to effectively manage and optimize various costs in production and operations. In market competition, cost advantages can be transformed into price advantages, attracting more consumers and increasing market share. Effective cost control also frees up funds for research and innovation.

Profitability is directly related to the survival and development of companies, indicating the ability to convert resources into profits. It generates economic value for companies and provides necessary financial support for innovation activities. Productivity, a combination of profitability and cost control, reflects the company’s ability to achieve higher output and efficiency with given resources. High productivity signifies the value added by each employee, which determines the resources available for innovation. Industrial robots enhance corporate competitiveness by reducing risk levels. Their highly accurate programming and execution capabilities significantly reduce human errors, leading to improved consistency and stability in product quality. According to cost-driving theory, their cost-control benefits enhance firms’ resilience in price competition and lower operational risks from rising factor costs. This enables firms to redirect more resources into R&D and product innovation, strengthening their overall risk-bearing capacity. According to reliability theory, the stable operation of industrial robots helps establish reliable production plans and delivery commitments, reducing customer-loss risks from production interruptions or delayed deliveries, and maintaining trust with downstream clients.

This paper measures the cost and profitability of a company in terms of operating cost ratio and profitability ratio, and measure company’s share of the market using its income ratio within the industry. Risk-taking as measured by the volatility of company’s earnings, and company productivity as measured by company’s total factor productivity (TFP) and labor productivity. Referring to Fu et al. (2021), we measure company productivity using both company labor productivity and TFP. Labor productivity is calculated by dividing the company’s value added by the number of employees. TFP is another measure of productivity of a company. The measurement of TFP is based on the Cobb-Douglas production function, which is quantified by estimating the fitted values and residuals through a two-step method, specifically including the OP method, LP method and OLS method. OP method and LP method are used to overcome the simultaneous and selection bias of labor and capital coefficient. The process of calculating the volatility of corporate earnings is shown in Appendix B^{Footnote 4}.

The results in columns (1) and (2) of Table 5 show that the coefficients of IBs are -0.976 and 0.129, respectively, and are statistically significant, indicating that IBs greatly reduce production and operation costs while significantly increasing profits. Column (3) shows that IBs significantly increases company’s market share. Columns (4) and (5) report the company’s risk-taking levels using standard deviation and extreme deviation measures, respectively, showing that IBs significantly reduces business risk. Table 6 demonstrates that the adoption of IBs improves TFP and labor productivity, as shown in columns (1) to (3). These results indicate that the use of IBs significantly enhances company productivity, regardless of the measurement method. Overall, these findings suggest that of IBs significantly improves cost control, profitability, market share, and productivity, while reducing business risks and enhancing company competitiveness. Combining these results with the benchmark regression, it can be inferred that companies innovate by improving its competitiveness through the use of IBs^{Footnote 5}.

Human capital inputs

Human capital is a crucial investment in innovation. IBs, as a prime example of automation technology, increase automation and change the human capital structure of companies, thereby affecting company’s innovation capabilities. The impact of automation technology on labor has long been a major concern in economics. The primary purpose of automation technology is to save labor by performing tasks that are repetitive, hazardous, or require low skills. Intuitively, as John Maynard Keynes suggested, technological unemployment occurs “due to our discovery of means of economizing the use of labor outrunning the pace at which we can find new uses for labor” (Autor and Dorn, 2013).

The replacement of human labor by robots is frequently perceived as a significant driver of unemployment in the labor market. However, a more optimistic perspective offers a counterpoint to this view. Since the Industrial Revolution, the ongoing development of labor-saving automated technologies has not resulted in widespread job elimination. Although automation and technology decrease the need for traditional labor in certain sectors, they simultaneously stimulate job creation by generating new demand for services and skills. This dynamic implies that while some job categories may decline, new opportunities emerge, fostering a more diversified and potentially more skilled labor (Leontief, 1983; Bessen, 2019). In addition, automation technology improves productivity, reduces the production costs (Herrendorf et al., 2013). Companies can produce more products and services with less labor (Autor, 2015). Automation technology enables companies to expand production scale, leading to increased employment opportunities (Koch et al., 2021). Aghion et al. (2022) highlight that the overall impact of automation on employment depends on the relative sizes of the substitution and creation effects on labor. Autor et al. (2003) propose a “task-based model” to analyze the substitutability of automation for different occupations by distinguishing between abstract and manual tasks, providing a foundational framework for examining the impact of automation technology on the labor market. Building on this, researchers have identified the phenomenon of job polarization, characterized by growth in high-education, high-wage tasks and low-education, low-wage tasks, alongside a decline in middle-wage, middle-skill tasks (Autor et al., 2006; Autor and Dorn, 2013; Acemoglu and Restrepo, 2020). Automation technologies tend to take replace middle-skilled labor performing routine, repetitive tasks, while increasing demand for both low-skilled and high-skilled jobs^{Footnote 6}. Educational attainment of employees plays a crucial role in the impact of automation. Highly educated workers generally perform fewer automated tasks than less educated workers (Arntz et al. 2016). Consequently, highly educated employees benefit from its adaptability to new technologies, whereas less educated employees face reduced employment opportunities (Zhou et al. 2020). Therefore, the skill bias induced by automation shifts company’s labor demand and the educational composition of the workforce, enhancing the company’s human capital (Acemoglu and Autor 2011). Additionally, human capital is a key factor of firm innovation. We examine the skill bias of IBs and its effect on the educational attainment of employees. Low-educated employees in manufacturing, typically involved in highly repetitive, easily standardized, and programmable tasks such as assembly, packaging, and quality control, are at greater risk of being replaced by IBs. Conversely, high-educated employees, particularly those with university degrees and above, are engaged in more innovative, managerial, and decision-making tasks requiring complex cognitive skills and creative thinking, which are not easily supplanted by robots. Acemoglu and Restrepo (2020) note that automation concurrently creates new jobs requiring high-end analytics and technical monitoring, thus increasing the demand for highly educated employees. Specialized skilled workers, such as trained technicians and engineers, possess the professional knowledge and skills crucial to the automation process. These workers are essential for maintaining and monitoring robot operations, as well as continuously optimizing production processes and technical equipment, thereby sustaining or even increasing the demand for this workforce. We analyzed the ratio of employees with university, junior college, and high school education within companies, where “university” includes undergraduates, masters, and PhDs, representing high-educated employees; “junior college” denotes employees with specialized skills in a particular field; and “high school” refers to employees with a high school education or less, representing low-educated employees. We then compare the impact of IBs on these three categories of labor.

Table 7 presents the impact of IBs on the labor structure, revealing a significant increase in the number of highly educated and professionally skilled employees in companies, alongside a notable decrease in the number of low-educated employees. This suggests that the integration of robots enhances the labor force structure and drives human capital upgrading. While robots are replacing humans in routine production tasks, they are concurrently creating more knowledge and technology-intensive roles that align with automation technology, such as robotics engineers, maintenance engineers, and data analysts. Highly educated and professionally skilled labor, as a form of human capital, serves as a critical driver of innovation. Its extensive knowledge and proficiency in technology application are essential resources for enhancing the innovation capacity of companies. The displacement of low-educated employees by robots, coupled with the absorption of highly educated and professionally skilled employees, underscores the enhancement of labor quality within companies, validating hypothesis H3^{Footnote 7}.

Capital input

Innovation in companies is inseparable from financial support. IBs can impact a company’s capital usage, typically leading to an increase in initial capital investment. However, this investment can enhance a company’s financial performance by improving production efficiency and reducing long-term operational costs (Duan et al., 2023). Based on these, this paper analyzes the sources of funding for company’s innovation activities, considering both external factors such as policy support and debt financing, as well as internal factors such as financial returns and liquidity constraints.

Government subsidies play a crucial role as an external source of funding for company’s innovation efforts (Gao et al., 2021). These subsidies effectively incentivize companies to engage in innovative activities by lowering companies R&D costs and risks, offering financial support, fostering collaboration between companies and research institutions, and providing targeted assistance for the advancement of specific technological domains (Li and Zheng, 2016). We systematically compile the data on government innovation subsidies by categorizing the subsidy amounts based on keywords found in the “government subsidy details” section of listed companies annual report notes.

The impact of IBs on company’s innovation financing is examined in Table 8. Column (1) reveals a significant increase in government subsidies for innovation due to the use of IBs. Government subsidies for innovation typically support projects with clear innovation objectives and potential economic benefits (Cheng et al., 2019), making companies with high IB penetration more likely to receive innovation funds.

Debt financing is another significant external funding source for companies, influencing the quality of innovation (Hall and Lerner, 2010). However, the role of debt financing in promoting corporate innovation is debated. While debt financing provides funds for innovation, it may also constrain companies from undertaking risky innovation activities due to risk aversion (Manso, 2011). Aghion et al. (2013) argue that companies require pressure to innovate, suggesting that a moderate level of debt may incentivize management to drive innovation. Therefore, total liabilities, short-term borrowings, and long-term borrowings are used as measures of corporate debt financing. Total liabilities, short-term borrowings, and long-term borrowings are selected as measures of corporate debt financing. Total liabilities reflect company’s overall external financing scale, while short-term borrowing addresses temporary funding gaps, and long-term borrowing supports fixed asset construction.

Columns (2)-(4) of Table 8 show the effect of IB use on external financing. The results indicate that IB use significantly reduces company’s total and short-term liabilities, particularly short-term debt. As IBs enhance productivity and profitability, company’s operating costs decrease, potentially increasing internal cash flow and reducing reliance on debt financing, especially short-term debt^{Footnote 8}.

Internal financial support is also crucial for innovation funding (Milani and Neumann 2022). The impact of IBs on company’s internal financial liquidity, measured by financial returns and cash flows, is analyzed^{Footnote 9}. Financial returns, measured as corporate investment returns minus operating profit, show no significant impact on innovation funding in column (5). However, in column (6), IBs significantly increase company’s cash holdings, easing liquidity constraints. This is attributed to IBs reducing operating costs, boosting profitability, and enhancing cash flow, providing stable financial support for innovation by optimizing operations and improving efficiency.

In our analysis, we examined the funding sources for company’s innovation resulting from IBs, considering both external and internal sources of funding. While the use of IBs shows no direct impact on the company’s debt financing or financial returns, it does lead to a significant increase in government subsidies for innovation and improves cash flow. This increase in subsidies and improved cash flow serves as an important source of funding for company’s innovation activities, supporting hypothesis H4.

R&D input

The preceding sections of this study have extensively examined the mechanisms through which IBs influence company’s human capital and financial aspects, ultimately promoting innovation. However, a more direct indicator of increased investment in company’s innovation endeavors is the rise in investment in R&D personnel and capital. The adoption of IBs enhances productivity and stimulates the labor demand in routine production tasks, thereby freeing up resources for innovation. Moreover, companies utilizing IBs tend to attract and retain skilled personnel. These companies, often leaders in their industries, are more inclined to invest in cutting-edge R&D projects, which in turn attracts highly skilled R&D personnel interested in technology and innovation. Consequently, the adoption of IBs spurs company’s investment in both R&D personnel and R&D capital, expediting their innovation processes.

The utilization of IBs markedly amplifies company’s need for R&D labor, as evidenced by the increase in R&D expenditures shown in columns (4)-(5) of Table 6. These outcomes clearly demonstrate that the incorporation of robotics streamlines production processes and motivates companies to enhance their technological development capabilities, supporting a broader spectrum of innovative activities through augmented personnel dedicated to R&D and increased R&D budgets.