What Pandemics Mean for Robots and Inequality

IMF Blog, April 19, 2021

Source details

Study focus: effect of past major pandemics on industrial robot adoption.
Pandemics analyzed: SARS in 2003, H1N1 in 2009, MERS in 2012, and Ebola in 2014.
Data and methods:
- Robot data from the International Federation of Robotics at the sectoral level.
- Coverage: 18 industries in 40 countries.
- Time period: between 2000 and 2018.
- Outcome measure for robot adoption: new robot installations per 1000 employees.
Additional country-level analysis linking robot adoption to inequality measured by the Gini coefficient.

Pandemic events are followed by increases in robot adoption (measured by new robot installations per 1000 employees).
The increase in robot adoption is especially pronounced when:
- the health impact of the pandemic is severe; and
- the pandemic is associated with a significant economic downturn.
Where new robot adoption has increased more following a pandemic, the medium-term rise in inequality (measured by the Gini coefficient) is larger.
Conclusion: acceleration of robotization is an important channel through which pandemics lead to higher inequality.

Firms restructure after large shocks (such as recessions) and adjust production toward technologies that lower labor costs.
Firms prefer robots because they are immune from health risks during pandemics.
Pandemic-induced uncertainty increases incentives for automation as firms try to withstand future pandemics.

Low-skilled workers are more at risk of displacement by robots than high-skilled workers.
The differential displacement reinforces existing inequality dynamics.
While automation and robotization are accelerating from still-low levels, they will likely become even more important drivers of inequality in the future.
Potential social consequences if disparities grow: long-lasting grievances and ultimately social unrest, producing a vicious cycle.

Policymakers should act to prevent scarring effects on the livelihoods of the most vulnerable, including through appropriate labor market policies.
Mitigation measures include:
- revamping education to meet demand for more flexible skill sets;
- lifelong learning and new training—especially targeted to the most affected workers (example cited: Singapore’s SkillsFuture initiative, which promotes learning in all stages of life).
Acknowledge limitations of training policies:
- Training that requires acquiring a substantively different and challenging set of skills may lead to dropouts.
- Therefore, policymakers should consider additional measures to address medium-term social challenges, including strengthened social safety nets.
Distributional outcomes of robotization depend on policy choices: a society willing to provide support to those left behind can accommodate a faster pace of innovation while ensuring broader wellbeing.

Source: IMF blog post “What Pandemics Mean for Robots and Inequality” by Tahsin Saadi Sedik and Jiae Yoo, April 19, 2021.