1/n Working paper w @AlexWeinberg17 and @pilossopher: Which workers bear the burden of social distancing policies? "Our main finding is that workers in low-work-from-home and high-physical-proximity jobs are more economically vulnerable across various measures".

2/n Main takeaways for policy: (i) lots of scope to target, (ii) high bang-for-buck! low bucks (affected are already low income), high bang (affected have lower liquid assets), (iii) double-edged sword: same, poor, people exposed to economic risk now + health risk on opening

3/n We use O*NET data to construct work-from-home (following @TradeDiversion and @BrentNeiman) and physical proximity at work measures and the OES to carefully cross-walk them over to individual data. This is how occupations look.

4/n We validate the measures by showing that they compare well with what we could glean from ATUS. Hopefully this is useful for other researchers that want validated measures. We'll have the data online shortly.

5/n We then look at worker characteristics across jobs. This is our main figure. It shows, for example, that workers in low work-from-home jobs are ~25ppt more likely to be below median wage, they are also 12pt more likely to have low liquid assets relative to income (in red)

6/n We then show that these pre-virus measures stack up in the Feb-Mar 2020 CPS. Low work-from-home occ employment contracted more, and the types that had large employment losses line up with our predictions. A 4ppt diff b/w low/high income (3.8m low inc jobs lost, 0.8m high inc)

7/n As further validation we show that MSAs with less employment in low work-from-home jobs experienced smaller increases in the fraction of phones that 'stay-at-home' between early-March and mid-April (Thanks @SafeGraph!)

8/n: Forgot... paper here: http://www.simonmongey.com/uploads/6/5/6/6/65665741/mongey_pilossoph_weinberg_work_from_home_april_2020.pdf