Job loss and Covid-19: an analysis on the impacts of remote work and automation

ABSTRACT

Using a unique dataset from a dedicated Cedefop Skills Forecast scenario on the impacts of COVID-19, this paper explores two possible determinants of expected job loss in the European Union (EU) due to the pandemic, namely the potential of work from home and the impacts of automation. Our findings suggest that less remote work and more automation are both related to future job losses across countries and occupations. These links are stronger in 2020–2021 at the country level, while becoming significant at the occupation level after 2022 when several protective measures taken by EU governments are expected to have been lifted.

KEYWORDS:

• COVID-19
• job loss
• home-based work
• automation risk
• Cedefop Skills Forecast

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Raw data were generated at Cedefop. Derived data supporting the findings of this study are available from the corresponding author [I.L.] on request.

Notes

¹ The sectoral WFH estimates are constructed as a weighted average of the occupational WFH estimates provided by Gottlieb et al. (2021), where the weights are the 2018 EU-wide within-sector employment shares of occupations. WFH at the country level was re-estimated using occupational level data weighted by the 2018 within-country occupational shares. The resulting index has is highly and statistically significantly correlated with the one provided by Dingel and Neiman (2020) (correlation coefficient 0.976).

² Cedefop Skills Forecasts is a dataset providing a comprehensive set of employment projections across sectors and occupations for all EU Member States plus a few more countries. Trends in future employment are estimated by using harmonized data and a single methodology to make results comparable across countries, sectors and occupations and are released every two years. The process involves individual country experts in peer reviewing and validating the results. For more details, see Cedefop, Eurofound (2018).

³ The assumptions were made at Member State level, using the most recent statistical and other data available at the time and concern several aspects of the Covid-19 impact including; (a) the lockdowns, such as the nature of the restrictions and the duration of travel restrictions, (b) labour market participation rates (using EU labour force survey data for the latest quarter available, and including the effects of short-time work schemes, absences, and temporary lay-offs on the 2020 average hours worked per week) (c) changes in aggregate demand, including impacts on consumer expenditure, investment and trade, and (d) government response measures, including working arrangements, fiscal support and any additional final expenditure measures implemented or announced by the time of the modelling exercise. For more details, see Cedefop (2021). The WFH amenability of different jobs and the potential effects of accelerating automation were not explicitly considered when constructing the scenario.

⁴ The results of these regressions should be interpreted with caution and their coefficient estimates should not be taken at face value, as the samples are not large enough to secure high degrees of freedom.

⁵ Namely highly-skilled non-manual, skilled non-manual, skilled manual and elementary occupations (ISCO 1-digit classifications 1 to 3, 4 and 5, 6 to 8, and 9 respectively).