1/n We've just completed a new study of the potential occupational exposure to Social Distancing. In other words, who is at greatest risk of job loss due to schools, government, business, churches and household compliance with CDC recommendations.
2/n The study involves matching OES occupational data with work context information from the O*NET data. The work context data tells us how much an occupation "works with others" and has "close physical proximity with others."
3/n These two factors are important, because some workers "work with others" online or through a phone. A call center would be an example.Likewise, some workers are in "close physical proximity" without interacting. Example might be an assembly line.
4/n It would take a combination of both for workers to be heavily exposed, and not be able to wear protective gear. The call center is obvious, unless work stations are close, while an assembly line worker is already wearing protective gear, to which more can be added.
5/n We use the highest context scores (ranging 1-100) to select occupations at risk of a large negative labor demand shock due to social distancing related to COVID-19. To that we add occupations associated with affected industries, despite their O*NET scores.
6/n This would include Commercial pilots. This caused us to reintroduce many types of workers with lower scores in one category or another, simply because their industry will be impacted.
7/n The result was 116 separate occupations, from which we eliminated likely public sector employees (school teachers, professors), since we are still working. We also eliminated healthcare and medical services workers who should anticipate growing demand, not job losses.
8/n We report these occupations in two tables. This is the restaurants, etc. industry and occupations. 
9/n This is transportation and retail.
10/n All told, this is some 28 million jobs, or about 17% of employment. The weighted average salary of these jobs is $32,774 (2018 OES data). So, these workers are clearly lower paid, on average than the typical worker. From these occupations we can also infer lower benefits.
11/N We also offered simulations with three scenarios over a 6 month period. The first example uses an 80% reduction ceiling which is similar to the CBO's excellent 2006 study. We then scaled it to 60% and 40%, from 1-6 months.
12/n These simulations offer two things. The first is the potential scale of long term Social Distancing on these workers. This is 17% of the labor force. The second is the composition of the most affected jobs. These are the workers who would benefit most from fiscal policy.
13/n This study uses O*NET data to flesh out already important discussions about appropriate fiscal policy recommended by @jasonfurman, @MichaelRStrain and @Austan_Goolsbee.
This is not a forecast, but it should offer some evidence of the scale and scope of economic risk.
This is not a forecast, but it should offer some evidence of the scale and scope of economic risk.
14/n Finally, the time for discussing this as a purely stimulative policy is past. UBI or other direct pay should be viewed as complementing public health policies designed to reduce the effects of Covid-19.
Study is here: bsu.edu/academics/cent…
Study is here: bsu.edu/academics/cent…
