This study explores optimal strategies in dealing with unobserved heterogeneity in the context of working conditions. A clusterwise strategy avoids stuck-in-the-middle solutions, and provides a better projection of the optimal solution into a low-dimensional space. This segment-level method identifies risk factors underlying working conditions as well as optimal allocation strategies using concomitant variables. Model calibration is performed at European level using survey data and provides important insights at country level.

--