Fear of technology-driven unemployment and its empirical basis

Debate over whether technological change replaces more jobs than was created before the 18th century. Nevertheless, this is being echoed in recent policy discussions (Mokyr et al. 2015). With 47% of U.S. jobs at high risk of automation in the near future (Frey and Osborne 2017), policymakers, managers and researchers are increasingly investing in helping to adapt to the labor market and social system for a time when most man-made work will be performed by machinery. . This trend has been accompanied by a wave of economic research about the interaction between technology, labor and economics.

Technological changes can affect labor through different channels

Both empirically and theoretically, the impact of technological change on labor can arise through multiple channels. Many technology devices are designed to save human labor by replacing workers. However, economic theory suggests that a number of compensatory mechanisms may unbalance the early labor-saving effects of new technologies (Acemoglu and Restrepo 2019, Baldwin et al. 2021). First, technological change can increase labor demand by creating new jobs directly linked to new technologies. In addition, technology-induced productivity releases increase production resources that can increase labor demand for other jobs within the same firm or industry.

Second, technology can increase labor demand by increasing consumer demand. This happens when new technologies increase productivity growth and, as a result, lead to lower production costs and lower consumer prices. In addition, new technologies can increase the marginal product of labor and capital, resulting in higher wages and a return on capital. The next two effects contribute to real income growth. If demand is resilient enough and responds positively to revenue growth and price declines, technologies can stimulate demand-induced output expansion (Besen 2020).

Systematically review experimental evidence on technology and employment

In a recent study (Hötte et al. 2022) we refer to these channels as (1) replacement, (2) restoration and (3) actual income effects and analyze their empirical basis through a systematic literature review.

We have identified 127 relevant studies that provide evidence of technological change and its impact on employment over the past four decades. In order to capture the impact of technological change at the border, we focus on the study of industrialized economics. Furthermore, in order to uncover the potential underlying differences, we pinpoint the measurements of five broad classification technologies that differ mainly from those studied in the literature: (1) Information and Communication Technology (ICT); (2) robot-expansion; (3) innovation survey; (4) productivity; And (5) a residual section containing various alternative indicators.

Labor-creation mutes the labor-replacement effect of technology

We find that the number of studies that support the effect of labor replacement is more offset than the number of studies that support the effect of labor creation / restoration and actual income (Figure 1, panel AC). This observation is reaffirmed when looking at studies that analyze the net employment impact of technological change, suggesting that the net impact of technology on labor should be positive rather than negative (Figure 1, panel D).

Figure 1 The division of the study according to the type of results reported for each effect tested

Formula: Calculations based on 127 studies collected from systematic literature reviews in Hötte et al. (2022).

Searches vary across technology groups

Investigations into five distinct technology sections show similar patterns, but with some subtle differences that deserve to be highlighted.

  • ICT: There is no evidence that the effect of replacement combines the effect of restoration and real income. However, the results suggest that restored jobs are qualitatively different from replaced jobs. The spread of ICT has largely positive employment effects for highly-skilled, non-routine, and service jobs.
  • Robot: We note that labor-saving effects are usually offset by robot-induced labor restoration. Unlike ICT studies, robot studies do not touch on the complementarity between robot and human labor in performing tasks. Thus, the labor-intensive effects of robots are probably related to the production, operation, and maintenance of such technologies.
  • Innovation: Studies that rely on innovation as a measure of technology often argue that the impact of employment depends on the type of innovation. Although product innovation has been shown to be largely labor-intensive, evidence of the employment impact of process innovation remains mixed.
  • Entertainment: When considering productivity improvements as a proxy for technological change, we find a fairly balanced balance between the number of empirical studies supporting substitution and the two labor-creation processes. Employment gains have largely been favorable for non-productive, high-efficiency and service work. These studies are often associated with theories that argue that technological change has led to structural changes in the growing process sectors and services from more primitive to the bottom of the supply chain, including redistribution of economic activity. Nevertheless, the effect of net employment observed in this study is not positive but negative.
  • Others: Finally, the results of studies relying on other / indirect systems of technology indicate that the effect of labor substitution is offset by the effect of labor-creation. The effects of employment have been mostly positive for non-productive labor, yet some studies have found positive employment effects for low-skilled workers, especially in service jobs.

Concerns about widespread technology-driven unemployment have no empirical basis

Although we find greater support for the labor-intensive effects of technological change, we are wary of the conclusion that technology has a positive net effect on employment. Nevertheless, we safely conclude that the impact of technology’s labor substitution is usually offset by a range of compensatory measures that lead to the creation or restoration of labor. Therefore, there does not seem to be an empirical basis for the fear of technology-driven widespread unemployment.

Although we have not found any strong evidence for quantitatively negative net employment impacts, the qualitative impact and distributional aspects of technological change on employment cannot be ignored. In particular, low-skilled manufacturing and manufacturing workers have been adversely affected by technological changes (Blanas et al. 2019). Therefore, effective upskilling and recycling strategies should be at the forefront of policy making. Many professions in both the lower and middle parts of the skill distribution will continue to evolve and will demand a modified set of skills due to technological advances. Therefore, the employment prospects of the relatively weaker groups can be significantly improved by investing in the right skill set. Even so, owning one is still beyond the reach of the average person. For these groups, targeted social support systems will be important.

Shortening and road ahead

To the best of our knowledge, we have the most comprehensive systematic literary review of the technology-labor relationship, and the first to carefully distinguish different technology and influence channels. This study provides an empirical basis for the political and scientific debate over the impact of technological change on the labor market.

However, although most of the empirical studies considered in this work cover the effects of widely disseminated technology, job opportunities that may be automated in the near future are constantly expanding (Brynjolfsson and McAfee 2014). Experimental evidence of the impact of artificial intelligence, quantum computing, virtual reality, biotechnology, nanotechnology, renewable energy and other emerging technologies that will soon affect our economy is limited.

In fact, no study in this column evaluates the impact of this new wave of technological innovation. To that end, it is unclear how much our results could be extrapolated in the future (Baldwin 2020); Real-time monitoring and ongoing research are needed to fully understand the emerging impact of the introduction of new technologies on the future of work.


Acemoğlu, D and P Restrepo (2019), “Automation and new work: how technology displaces and restores labor”, Journal of Economic Prospects 33 (2): 3–30.

Baldwin, R. (2019), The Rise of Globotics: Globalization, Robotics and the Future of JobsOxford University Press.

Baldwin, R, JI Haaland, and AJ Venables (2021), “Jobs and Technology in a General Balance: A Three-Resilient Approach”, CEPR Discussion Paper 15739.

Basen, J (2020), “Automation and Jobs: When Technology Increases Employment”, Economic policy 34 (100): 589–626.

Blanas, S, G Gancia, and SY Lee (2019), “Machines and workers: how different technologies affect different workers”, VoxEU.org, 10 October.

Brynjolfsson, E and A McAfee (2014), The Second Machine Age: Work, Progress and Prosperity in the Age of Bright TechnologyWW Norton & Company.

Frey, CB and MA Osborne (2017), “The Future of Employment: How Sensitive Are Jobs to Computerization?”, Technological forecasting and social change 114: 254–280.

Hötte, K, M Somers, and A Theodorakopoulos (2022), “Technology and Jobs: A Systematic Literary Review”, Oxford Martin Working Paper Series on Technological and Economic Change No. 2022-2.

Mokyr, J, C Vickers, and NL Ziebarth (2015), “History of Technological Concerns and the Future of Economic Growth: Is This Time Different?”, Journal of Economic Prospects 29 (3): 31-50.

Leave a Reply

Your email address will not be published.