Robotic arms along assembly line in modern factory. (Image: photocreo via Envato)

What is driving the adoption of robots in industry?

Novel plant-level data from Germany sheds light on the current status and the potential determinants of robot use in manufacturing factories.

By Liuchun Deng, Verena Plümpe and Jens Stegmaier

The unprecedented rate of advances in automation technology during the last decade has spurred intense debates about the future of globalization, labour and society at large. The rise of industrial and service robots is central to this discussion.1 Unlike the textile machinery destroyed by the Luddites in the 19th century, modern robots are (re)programmable, multipurpose and increasingly capable of performing a range of tasks without human intervention. Equipped with artificial intelligence, robots are expected to assume a major role in the work environment beyond the traditional factory setting.

The evolution of the use of robots in Germany

To gain further insights into the rise of robots, we analysed a newly collected, plant-level dataset on robot use in Germany.2 To our knowledge, it is the first empirical study that leverages a firm-level dataset to study robotization in Germany3, a country renowned for its production of and innovation in robots.4 In what follows, we describe the data used in our study, present key stylized facts on recent robotization in Germany, and provide evidence on potential determinants of robot adoption.

The data on robots was derived from the IAB Establishment Panel, a high-quality, representative annual survey of around 16,000 German plants. We included a section specifically on robots in the 2019 wave, asking respondents (1) whether their plant had used robots between 2014 and 2018; if so, (2a) the number of robots used during each year from 2014 to 2018, and (2b) the number of new robots purchased in 2018; and (3) the characteristics of the installed robots. The resulting dataset is the first longitudinal dataset of direct robot measures for Germany at the plant level.

Robot use is rare and skewed towards larger plants

Robot use at the plant level is still relatively uncommon, even in a country with the highest robot density in Europe. In 2018, only 1.6 per cent of German plants used robots. In the manufacturing sector, where industrial robots are presumed to be used more heavily, robot users accounted for only 8.2 per cent of the plants in 2018, while in the non-manufacturing sector, 0.9 per cent of the plants had installed robots. There is substantial heterogeneity in the share of robot users across industries within the manufacturing sector, ranging from 25.6 per cent (plastics) to 3.3 per cent (textiles).

Robot distribution in manufacturing sector in 2018

Notes: (1) Plants are sorted by the number of robots reported in 2018. (2) The survey weights are applied. (3) Average robot count is rounded to the closest integer.

Source: Deng et al (2020).

Robot distribution in manufacturing sector in 2018 (top decile)

Notes: (1) Plants are sorted by the number of robots reported in 2018. (2) The survey weights are applied. (3) Average robot count is rounded to the closest integer.

Source: Deng et al (2020).

In 2018, 52 per cent of Germany’s entire robot stock was installed in 5 per cent of robot-using plants. The manufacturing plants in the top decile ranked by robot count had 40 installed robots, on average, compared with the median robot user, which had only two installed robots. Within the top decile, robot distribution remains highly skewed: the highest two percentiles had 141 installed robots, on average. It should be noted that this skewed distribution of robots is not just reminiscent of employment distribution. In fact, robot density in the top decile (deciles defined as in the first figure above), measured by the number of robots per 1,000 workers, was about 10 times higher than that of the median user.

Share of robot users: 2014 vs. 2018

Notes: (1) A plant is identified as a robot user in 2018 if it answered yes to the question of whether it used robots from 2014 to 2018 and its robot stock in 2018 was not zero. (2) Survey weights in 2018 are applied. (3) The share of robot users in 2014 is estimated based on the survey answers to robot use in 2014 and 2018.

Source: Deng et al (2020).

More plants have started adopting robots

Many plants installed robots for the first time between 2014 to 2018. The new robot users contributed substantially to robotization in Germany. The figure above compares the share of robot users by industry in 2014 and in 2018. The share of robot users in the manufacturing sector increased by more than half from 5.2 per cent to 8.2 per cent. Similarly, the share of robot users in the non-manufacturing sector increased significantly from 0.5 per cent to 0.9 per cent. The figure below decomposes the growth of the robot stock by industry between 2014 and 2018 into the contribution of existing users in 2014 and of new users that adopted robots for the first time between 2014 and 2018. The installation of new robot by users accounted for a considerable share of growth in robot stock. 

Decomposition of growth in robot stock: new vs. existing users

Notes: (1) Calculations are based on the surveyed plants that reported their robot use in each year from 2014 to 2018. Survey weights in 2018 are applied. (2) For each sector, the contribution of the robot adopters to growth is defined as the ratio of the total robot stock of robot adopters in 2018 to the robot stock aggregated over the existing users in 2014. The contribution of the robot users to growth is defined as the percentage change of the aggregate robot stock from 2014 to 2018 for the plants that already used robots in 2014.

Source: Deng et al (2020).

Robot users are still not the rule, but the exception

A number of plant-level characteristics differentiate robot users from non-users. According to our estimation, robot users employ around four times as many employees as non-users. For plants with the same level of employment, robot users tend to hire more low-skilled labour, are more likely to export, to adopt new technologies, and have a higher level of investment. Among the sample of robot users, plants that install more robots also have a higher level of employment.

The rise of “cobots”, i.e. robots that can collaborate with human workers, is shaking up the stereotype of robots that have traditionally been installed along production lines. In 2018, only 49 per cent of German robot users reported that they were using robots that operated independently of employees with the help of a protective device ( “cage robots”), which differ from the new collaborative robots. Robotization premia are more pronounced among cage robot users. Compared with other robot users, cage robot users hire more employees, have higher labour productivity, and are more likely to engage in export.

Potential determinants of robot adoption

We examined potential determinants of robot adoption, which could help explain whether robotization premia already existed prior to robot adoption. Following the existing theoretical framework of robotization5, our exploration focused on five firm-level aspects: plant size, productivity, share of low-skilled labour, exporter status and labour costs.

Our empirical analysis revealed a strong self-selection in robot adoption. The likelihood of a plant newly adopting robots during the period 2015 to 2018 strongly depended on the plant’s characteristics in 2014. Specifically, we found that plant size, share of low-skilled labour and exporter status are positively associated with subsequent robot adoption. The positive effects are statistically and economically significant.6 Once we control for initial plant size, however, productivity measures seem to have little, if not negative, effects on robotization.

Germany introduced a uniform minimum wage on 1 January 2015. Exploiting this significant change in policy on labour cost, we further documented that manufacturing plants affected by the minimum wage regulation were more likely to adopt robots. This is consistent with the theoretical prediction that when labour costs rise plants have more incentives to automate.7

Concluding remarks

The stylized facts that we present here underscore the importance of examining the rise of robots at the plant level. Even though our analysis focused exclusively on Germany, a comprehensive picture of robotization will emerge in the near future with the increasing availability of plant- or firm-level data on robots across countries.8

Our investigation of the determinants of robot adoption highlights how both external factors, such as participation in international trade, and internal policies, such as the introduction of minimum wage, influence plants’ decision to adopt robots. The findings shed light on the drivers of automation and thus inform the ongoing policy-relevant discussion.

  • Liuchun Deng is Assistant Professor of Economics at Yale-NUS College and Research Affiliate at the Halle Institute for Economic Research (IWH).
  • Verena Plümpe is PhD student in the Department of Structural Change and Productivity at the Halle Institute for Economic Research (IWH).
  • Jens Stegmaier is Senior Researcher at the Institute for Employment Research (IAB) in Nuremberg and Research Affiliate at the Halle Institute for Economic Research (IWH).

Disclaimer: The views expressed in this article are those of the authors based on their experience and on prior research and do not necessarily reflect the views of UNIDO (read more).


  1. Baldwin, R (2019), The Globotics Upheaval: Globalization, Robotics, and the Future of Work, Oxford University Press. For the empirical work on robots, see the seminal paper: Graetz, G and G Michaels (2018), “Robots at work”, Review of Economics and Statistics 100(5): 753-768.
  2. Deng, L, V Plümpe, and J Stegmaier (2020), “Robot adoption at German plants”, IWH Discussion Paper.
  3. Earlier version of the article was published on VoxEu. (2020) Robot adoption at German plants. 16 January 2021.
  4. For earlier work using the industry-level robot data to study the effects of robotization in Germany, see Dauth, W, S Findeisen, J Südekum, and N Wößner (2020), “The adjustment of labor markets to robots”, Journal of the European Economic Association, forthcoming.
  5. Acemoglu, D and P Restrepo (2020), “The race between man and machine: Implications of technology for growth, factor shares, and employment”, The American Economic Review 108(6): 1488-1542.
  6. What we have documented echoes findings using the Spanish firm-level data in Koch, M, L Manuylov, and M Smolka (2019), “Robots and firms”, The Economic Journal, forthcoming.
  7. For a similar finding using the Chinese data, see Fan, H, Y Hu, and L Tang (2020), “Labor costs and the adoption of robots in China”, Journal of Economic Behavior and Organization, forthcoming.
  8. See, for example, Bonfiglioli, A, R Crino, H Fadinger, and G Gancia (2020), “Robot imports and firm-level outcomes: Evidence from French firms”, Working Paper; Humlum, A (2019), “Robot adoption and labor market dynamics”, Working Paper.

Read next