This talk is co-sponsored by Rutgers Global. It is open to the public, but registration is required. Click here to register.
The influx of industrial robots into China has witnessed a remarkable increase since early 2000s. According to reports by International Federation of Robotics (IFR), China’s share of global industrial robot installations was only 0.34% in 2000, with a mere 380 units installed. This figure skyrocketed to 12.4% in 2010 and surged to nearly half of the global total in 2020 (44%), with 168,337 unites installed within China. Among the various manufacturing sectors in China, the automotive and electronic industries lead the charge in important industrial robots.
Industrial robots are poised to impact firm productivity by streamlining automations, improving performance, increasing labor productivity, facilitating knowledge transfer, and accelerating the learning-by-doing process. It is worth noting that the majority of existing literature predominantly focuses on developed countries, with limited attention given to developing nations. This is largely due to the delayed adoption of industrial robots in the latter and the scarcity of relevant data. Furthermore, many of these studies rely on macro-level data, while only a few delve into the effect of industrial robots on firm performance using firm-level data. The objective of this study is to address the gaps in the literature by examining the factor influencing the adoption of robots and assessing the effects of industrial robots on productivity within manufacturing firms in China.
This study utilizes two longitudinal databases. First, the Chinese Annual Surveys of Industrial Firms (ASIF), conducted by the National Bureau of Statistics, provide firm-level information. This dataset covers more than 90% of industrial output and over 95% of exports in China. ASIF enables us to estimate firm productivity. The second dataset is compiled from customs import data provided by the Chinese General Administration of Customs (CGAS). Merging the two datasets yields a total of 2,100,035 manufacturing firms spanning 15 industrial sectors in 2000-2013, with 920 firms have imported industrial robots during this period.
This study provides several noteworthy findings. First, firms were more inclined to adopt industrial robots when they exhibited higher productivity, possessed a larger market share, engaged in export activities, had a shorter establishment tenure, and demonstrated financial robustness. Furthermore, foreign or joined ventures and state-owned enterprises showed a higher likelihood of adopting robots. Additionally, an increase of labor cost, as measured by the average wage paid by the firm and the provincial minimum wage, was found to facilitate the adoption of robots. Interestingly, this study does not reveal a significant impact of the dependency ratio for elderly and youth on the likelihood of adopting robots.
Second, based on the propensity score matching and difference-in-difference (PSM-DID) estimation, we find a statistically significant and positive casual effect of industrial adoption on firm productivity, resulting in an approximate 10% increase in total factor productivity. These effects were more pronounced in the first two years following adoption but lost statistical significance from the third year onward. Moreover, we find that the adoption of robots at the city level also generated positive spill-over effects to non-adopters within the same industry, leading to productivity improvements for these firms as well. In addition, the effects of robots on firm productivity varied by robot types and industry. Specifically, single-functional robots exhibited a more substantial effect compared to multi-functional robots; and the automotive and chemical industries experienced greater productivity gains from robot adoption compared to the electronical, electric, and equipment industries. These findings remained robust across various checks, including assessments of unobservable selection bias as well as using alternative matching algorithms, outcome variables, robot measurements, and estimation approaches, underscoring the validity and validity of our findings.
This paper makes several significant contributions to the existing literature. First, it offers valuable insights into the impact of robot adoption on firm productivity within the context of China, shedding light on the implications for developing countries. Second, this study stands out as the first to address selection biases on both observables and unobservables within the literature on robots. Third, it is also the first study to distinguish between single- and multi-functional robots and examine their differentiated effects on firm productivity. Finally, it addresses a specific data issue concerning Foxconn Electronics and its affiliated companies, which deployed their own robots in 2000-2013 when most manufacturing companies in China relied on imported robots. This data concern has been overlooked in the existing literature.
Yanhong Jin is a professor in the Department of Agricultural, Food and Resource Economics in the School of Environmental and Biological Sciences (SEBS) at Rutgers University.