Suitability for Machine Learning and Immunity to the COVID-19 Pandemic
52 Pages Posted: 10 Jul 2020 Last revised: 9 Dec 2021
Date Written: July 25, 2020
Abstract
We evaluate economic immunity to the COVID-19 pandemic accorded by a sector's suitability for machine learning (SML). We use an event study design, which causally estimates the impact of sectoral SML on variation in economic activity prior to and following the first pandemic-induced lockdown in India. Using payments data from one of India's largest payment gateways, we find a one-standard-deviation increase in SML arrests the pandemic engendered decline in sectoral payments by a third. High SML of downstream sectors further improves a sector's economic resilience, emphasizing the importance of a key driver of the impact of firms' strategic technology choices: the ability of downstream value chain partners to embrace technological change. We provide suggestive evidence of two mechanisms underlying the observed economic resilience, notably, greater business continuity and adaptability to remote work along the SML continuum. Textual analyses of tweeting activity of firms suggest that firms with high SML continue to tweet similar content prior to and following the lockdown. Employee-level surveys reveal that high SML relates to pre-pandemic institutional knowledge and infrastructure as well as greater incidence of remote work during the lockdown.
Keywords: Automation, SML, Economic Activity, Remote Work
JEL Classification: E23, E24, J01, J22
Suggested Citation: Suggested Citation