How do you combine new and old data to predict which jobs are at risk from social distancing?

September 25, 2020

Whose job is most at risk from the C-19 economic downturn? How much subsidy would compensate different types of businesses for the recent social distancing measures? How can you make forecasts when official statistical data collection is simply too slow? CEU’s Miklos Koren and Research Fellow Rita Peto (The Centre for Economic and Regional Studies) address these questions in their latest research (https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0239113), published in the prestigious international journal PLoS ONE. As Koren, Professor in CEU’s Department of Economics and Business notes, this type of research “offers an excellent example of how combining economic theory, traditional statistics, and newer data sources can help better understand the crisis resulting from the pandemic, and therefore help predict economic trends.”

By May 2020, one in eight of the more than 110 million workers in the United States had lost their jobs as a result of C-19. The biggest layoffs were in sectors that had come to an almost complete halt: restaurants closed, people stopped travelling and, with the exception of basic food and medicine, the number of shopping trips also fell drastically.

Koren and Peto relied on large-sample survey data to detect the most vulnerable sectors where personal communication with customers was paramount. These sectors are retail stores, arts, entertainment, and recreation, hotels and restaurants, where more than 38% of employees have jobs that require frequent personal communication. The exposure of these sectors to C-19 is well known, but this research highlights the magnitude of the problem for workers.

In addition, the authors relied on data supplied by business analytics firm SafeGraph to capture, in real time, the dramatic fall in in-person commercial transactions. This data comes from anonymized location information from several smartphone apps. The fall in customer visits was indeed largest in vulnerable sectors, where face-to-face contact with customers is important. The two effects magnifying one another lead to a large number of workers being laid off or furloughed. In the most vulnerable sectors, the study estimates that businesses would need about 126% wage subsidy for each of their workers.