By: Pablo A. Egaña del Sol (GSAS’16)
Date: July 2020
Field: Economics, Public Policy
Special thanks to: Gabriel Cruz (UAI)
For more information: Pablo.email@example.com
COVID-19’s Impact on the Labor Market shaped by Automation: Evidence from Chile
Impactos del COVID-19 Modulados por la Automatización en el Mercado Laboral: El Caso de Chile
In this article we argue that the ongoing COVID-19 pandemic would serve as a catalyzer for automation processes in several firms across industries. We consider Chile, a small, open developing economy, as a case study. It is a desolate fact that many workers lost their jobs in recent months due the restrictions (quarantine and confinements, curfews, general movement restrictions, among others) put in place to control the COVID-19 pandemic. Many of these workers are expecting to return to their jobs when the conditions turn out to be safer and the government lifts restrictions. Unfortunately, this may not be the case.
Even though it is a very complicated, rather speculative, task to predict the situation in the labor market after the COVID-19 pandemic, we present some facts that are consistent with the following hypothesis: industries are accelerating the digital transformation of their operations and, as a consequence, Chile might experience a jobless recovery in many sectors, especially in those where automation technologies are available, the degree of at-work physical proximity is high, the level of exposure of infectious diseases, such as COVID-19, is high, and the possibilities to work remotely are low. In other words, several firms either are or will be forced to operate relying heavily on labor-saving technologies during the pandemic. For instance, agriculture and domestic service sectors are the most automatable areas and their possibilities to work remotely is very low. Furthermore, domestic services have a high level of exposure to diseases. In this sense, domestic service is a sector that may suffer high consequences due to COVID-19 and the automation process.
In this context, this project has two objectives: (1) Identify which industries, territories and economic activities will be most impacted by automation due to COVID-19; and (2) measure the effects that COVID-19 has had in Chile’s labor market. For logical reasons, the latter will be assessed after we overcome the pandemic.
We provide empirical evidence supporting our hypothesis using data at individual and industry/sector level in Chile. In particular, we use data from Encuesta de Caracterización Socioeconómica Nacional (CASEN) in order to get information about workers' occupations, age, gender, geographical area, and other relevant characteristics. We merged this data with two different datasets containing information about the risk and the degree of exposure to automation at occupation level, which is based on expert assessments (Frey & Osborne, 2017) and patent records (Webb, 2020), respectively. Finally, we included another source of information related to the degree of physical proximity, the level of exposure of infectious diseases (i.e. COVID-19) at the workplace, and the capabilities possibilities rather than capabilities to work remotely (Beland et al., 2020). Using this novel dataset, we were able to predict the degree of automation risk by worker’s demographics, occupations, industries, territories and economic activities in Chile. The methodology is described in greater detail here.
We find evidence that COVID-19 might act as catalyzer for the digital transformation of firms in Chile. In particular, we find that the automation risks have a significant geographical heterogeneity, even within subnational areas. This heterogeneity is partly determined by economic activities - e.g. agriculture, mining, and services - as well as by demographic characteristics (i.e. worker’s average age, income, and gender). For instance, 13% of female workers, compared to 21% of their male colleagues, are under high risk of automation. Furthermore, the lowest income quintiles have higher risk of automation. A clear example is possible to observe in the north. Antofagasta province has a 0.06 fraction of its workers under high risk of automation, while in Tocopilla province it is around 0.35, both according to Webb´s index of automation risk. Antofagasta is a service-oriented city for the whole North macrozone, while Tocopilla has an economy based on mining, energy and fishing (and particular, fishmeal production), which, as we present in more details in the document, has a larger probability to be affected by different array of automation technologies.