Fabrizio Colella (USI Lugano)
Max Bres, David Koll, Philipp Kircher and Vinzenz Ziesemer
This paper develops a forecasting algorithm for the occupational structure of the labor market. Using French administrative data from 1997–2006, it identifies leading firms—those ahead of aggregate trends—through a machine learning approach. A simple forecast equation incorporating the occupational structures of these firms significantly outperforms benchmark models in out-of-sample predictions for 2002–2006.