Early in the COVID-19 pandemic, it quickly became apparent that initial job losses from the ensuing economic crisis tended to be concentrated among low-wage workers (Cajner et al. 2020), women (Alon et al. 2020), and members of some minority groups (Fairlie et al. 2020). With the promise of an effective vaccine becoming widely available in 2021, attention turns to how these hard-hit demographic groups will fare in the post-COVID-19 economy.
Autor and Reynolds (2020) make a compelling case that the dwindling employment prospects currently facing low-wage workers may persist after the pandemic ends. They point out that the continued use of telepresence technologies by former commuters and business travellers may permanently decrease demand for many in-person low-wage service occupations. They also note that COVID-19 restrictions have forced automation in many industries and argue that many firms will not revert to their pre-pandemic production process even after the workplace is safe to return to.
There is concern that COVID-induced automation may increase inequality by exacerbating what Acemoğlu (2020) refers to as an “unbalanced technological portfolio”. Acemoğlu and Restrepo (2017) find that the increased use of robots in recent decades reduced per capita employment in the US, with workers with education less than college and those in routine manual occupations suffering the largest losses.
In a recent paper (Chernoff and Warman 2020), we use information from the Occupational Information Network (O*NET) to construct indexes of automation and viral transmission risk. The conceptual link between these two indexes is that pandemic risk may incentivise firms to automate tasks previously completed by workers. We define high-risk occupations as those with automation potential and transmission risk indexes both greater than or equal to 0.5 (both indexes are normalised to range between zero and one).
Using data from the American Community Survey, we identify the US local labour markets and demographic groups that may be most impacted by the possible push to automate jobs due to an overlap in viral transmission risk and automation potential. We also examine the distribution of high-risk occupations across demographic groups for 25 other countries using data from the Programme for the International Assessment of Adult Competencies (PIAAC).
Figure 1 suggests that there is no well-defined geographical pattern in the US with regards to the distribution of high-risk occupations. Instead, we find important demographic differences. We uncover that women in the US are about twice as likely as men to be in high-risk occupations.
Figure 1 highlights that women with less than a bachelor’s degree stand out as being at highest risk of both transmission and automation. When we further disaggregate by earnings, race, and education, we discover that this risk is always higher for females relative to males in the same group.
Figure 1 In the US, women with less than a bachelor's degree are at highest risk of COVID-19-induced automation
Notes: Legend values correspond to the fraction of the commuting zone population with automation potential and transmission risk indexes both greater than or equal to 0.5. Automation potential and transmission risk indexes are created from the O*NET and normalised to range between zero and one. Estimates are from the weighted counts from the 2013 to 2017 American Community Survey. The sample is restricted to individuals between age 18 and 65.
While our paper characterises the risks stemming from an overlap of automation potential and viral transmission risk, the results of previous work looking at these phenomena in isolation corroborates our findings. For example, Blanas et al. (2020) find that the fall in demand resulting from automation is felt strongest by low- and medium-skill workers as well as women. Baylis et al. (2020) suggest that the higher unemployment experienced by women early in the COVID-19-pandemic may be partially attributed to their heavy representation in occupations with a higher risk of viral transmission.
We find very similar results in other countries. In all 26 countries in our analysis, we find a greater fraction of females than males in high-risk occupations. As seen in Figure 2, workers with less than a bachelor’s degree are more likely to be in occupations with high transmission risk and automation potential in nearly all countries.
Figure 2 The higher risk facing less-educated women is common across countries
Notes: The horizontal bars measure the fraction of the population that work in occupations with automation potential and transmission risk indexes both greater than 0.5 (both indexes are normalised to range between zero and one). The US values are calculated using data from the American Community Survey and O*NET. Values for all other countries are calculated using data from the Programme for the International Assessment of Adult Competencies and O*NET. The sample is restricted to individuals between age 18 and 65.
While this is true for both sexes, the pattern is driven primarily by the high-risk occupations held by women with low- to mid-level educational attainment. Our international analysis also examines the differences in risk by wage level. As with the US, we find that occupations held by females with mid to low-level wages face the highest risk.
Bessen (2019) shows that there are many historical examples where automation was accompanied by employment growth rather than massive job losses. Similarly, COVID-19-induced automation may benefit some workers as wages may increase in occupations where new technologies complement workers’ skills. However, workers may also be displaced and face earnings losses in occupations where jobs are readily automated and viral transmission is high.
Our findings suggest that COVID-19-induced automation may exacerbate labour market disparities, as females with mid to low levels of wages and education appear to be at the highest risk of being negatively affected.
Authors’ note: The views in this column are those of the authors and do not necessarily reflect those of the Bank of Canada.
References
Acemoğlu, D, and P Restrepo (2017), “Robots and jobs: Evidence from the US”, VoxEU.org, 10 April.
Acemoğlu, D (2020), “Machines, artificial intelligence, and the workforce: Recovering and readying our economy for the future”, US House Committee on the Budget Hearing, witness testimony, 10 September.
Autor, D, and E Reynolds (2020), “The nature of work after the COVID crisis: Too few low-wage jobs”, Brookings Institute.
Alon, T, M Doepke, J Olmstead-Rumsey and M Tertilt (2020), “The shecession (she-recession) of 2020: Causes and consequences”, VoxEU.org, 22 September.
Baylis, P, P-L Beauregard, M Connoll et al. (2020), “The distribution of COVID-19 related risks”, NBER Working Paper 27881.
Bessen, J (2019), “Automation and jobs: When technology boosts employment”, VoxEU.org, 12 September.
Blanas, S, G Gancia and S Yoon (Tim) Lee (2020), “Who is afraid of machines?” Economic Policy 34(100): 627–90.
Cajner, T, L D Crane, R A Decker, J Grigsby, A Hamins-Puertolas, E Hurst, C Kurz and A Yildirmaz (2020), “The US labor market during the beginning of the pandemic recession”, NBER Working Paper 27159.
Chernoff, A, and C Warman (2020), “COVID-19 and implications for automation”, NBER Working Paper 27249.
Fairlie, R W, K Couch and H Xu (2020), “The impacts of COVID-19 on minority unemployment: First evidence from April 2020 CPS microdata”, NBER Working Paper 27246.
