Measuring Working From Home Adoption: Evidence from half-a-billion online job ads
Raffaella Sadun (Harvard); Nick Bloom (Stanford); Steven Davies (Chicago); Stephen Hansen (Imperial); Peter Lambert (LSE); Bledi Taska (Burning Glass)
Abstract
We use novel data on half a billion job vacancies across several countries to study the adoption of working from home before and after the Covid pandemic. We leverage novel machine-learning techniques to detect heterogeneity in work-from-home features across jobs, and job characteristics that are typically associated with WFH adoption. The study reveals significant heterogeneity in adoption across and within countries. We match the data with firm-level information and document the extent to which the variation can be accounted for by idiosyncratic firm-level policies vs. country-specific factors.