Essays on Machine Learning and Labor Economics
Abstract
Observed worker and firm characteristics only explain a small wage variation. Beyond characteristics that are directly observed from the data, my thesis develops new empirical methods aimed at identifying unobserved heterogeneity in the labor market. Chapter 1 proposes an empirical method to measure the effects of coworkers on wages. I take advantage of the recent cutting-edge clustering method that combines machine-learning and economic theory to identify groups of workers with similar latent productivity type. I further apply the cluster-based method to identify the effects of coworkers on wages and evaluate their economic implications in empirical-relevant simulations. The proposed method has proven potential to be applied to the real-world data to improve our ability to understand the role of coworkers in substantive questions where existing methods have limitations.
Subject Area
Labor economics|Economics
Recommended Citation
Xin, Jianhong, "Essays on Machine Learning and Labor Economics" (2022). Dissertations available from ProQuest. AAI29321392.
https://repository.upenn.edu/dissertations/AAI29321392