Is the Covid-19 pandemic fast-tracking automation in developing countries? Evidence from Colombia

Leonardo Bonilla; Luz A Flórez; Didier Hermida; Francisco Lasso; Leonardo Fabio Morales; Juan Jose Ospina; José Pulido

Is the Covid-19 pandemic fast-tracking automation in developing countries? Evidence from Colombia

BIS Working Papers | No 1048 |

10 November 2022

by Leonardo Bonilla, Luz A Flórez, Didier Hermida, Francisco Lasso, Leonardo Fabio Morales, Juan Jose Ospina and José Pulido

PDF full text

(1,004kb)

| 34 pages

Summary

Focus

Recent evidence suggests that the Covid-19 pandemic has tended to accelerate the automation process in developed economies. We assess whether this is also happening in Colombia, a developing country characterised by a combination of low investment in research and development (R&D) and productivity, and high levels of informal labour and unemployment. In this context, the mobility restrictions imposed by the government in order to reduce the spread of the virus and the fear of contagion led to a reduction in labour supply, increasing labour cost relative to capital, which in turn could have enhanced automation.

Contribution

We contribute to the literature about the pandemic's impact on occupations more prone to automation, providing evidence from a middle-income country like Colombia with high levels of informal labour and unemployment. We measure the demand for new jobs during the pandemic using vacancies by occupations collected by the Colombian Public Employment Services Bureau and the total salaried employment level based on household surveys. The probability of automation of each occupation is measured using the Frey and Osborne (2017) and Nedelkoska and Quintini (2018) methodology adapted to the Colombian case. We evaluate the differential effect of the pandemic on job openings and occupations using event-study methods

Findings

During the pandemic, job openings fell more drastically in occupations with a higher potential of automation. These effects are large and persistent, with negative and significant coefficients until our last period of study (August 2021). We also find negative and significant effects on employment, particularly salaried employment. Most of our results are driven by the sectors that were more affected by mobility restrictions. Therefore, in these cases, mobility restrictions amplified the effect of the pandemic on automation. These results are robust to alternative specifications, such as using different measures of automation potential or controlling for sector-specific trends. Finally, we show that the most affected individuals are women over 40 who work in low-wage occupations.

Abstract

This paper assesses whether the Covid-19 pandemic accelerated automation in developing countries. We studied the case of Colombia, a country with low R&D and productivity and with high labor informality and unemployment. We estimated event-study models to assess the differential effect of the pandemic on job openings and salaried employment by the potential degree of automation of each occupation. Our results suggest that both vacancies and salaried employment fell more in highly automatable occupations during the pandemic and have since experienced a slower recovery. The effect of the pandemic on automation is mostly driven by sectors that were affected by mobility restrictions. We also found heterogeneous effects by age and gender. The acceleration of automation is mainly affecting the labor market for females and individuals over the age of 40. Finally, we explored the differential effect on occupations with wages around the minimum wage. We found that occupations with wages close to the minimum wage exhibit the highest effect, especially at the onset of the pandemic.