"JAQ of All Trades: Job Mismatch and Firm Productivity" with Luca Coraggio, Marco Pagano and Joacim Tag. 

Abstract:

Using machine learning, we predict workers’ suitability to jobs, based on the observed job assignment in the most productive firms and conditioning on workers’ characteristics and their complete work history. This enables us to construct a firm-level index of job assignment quality (JAQ), which we validate by showing that it is positively associated with firm productivity. We further analyze whether JAQ increases around major organizational changes.