Joint work with Alonso Alfaro-Ureña and Jose Vasquez. Extended abstract (for conference submissions) available on request.

Abstract. We document an ample degree of dispersion of supplier quality, defined as the effect of a supplier’s inputs on its buyer’s sales, in the Costa Rican production network. Supplier quality also appears uncorrelated with buyer unobservables, suggesting the existence of both informational and spatial frictions affecting firms’ choice of suppliers. We quantify these two forces via a structural model of production network formation.

Joint work with Yaroslav Korobka. Extended abstract (for conference submissions) available on request.

Abstract. Borrowing tools from the practice of neural networks, I design an empirical framework for the analysis of “hierarchical networks:” socio-economic settings featuring multiple, layered networks, whose nodes are linked across layers. I use this framework to revisit questions involving networks of workers and fi rms.

Joint work with Aslan Bakirov and Francesco Del Prato. Draft coming soon.

Abstract. We revisit the wage decomposition literature using machine learning. We show empirically that if both worker- and firm-level observable characteristics are treated non-parametrically via generalized random forests, the share of log-wages variance explained by typical “AKM” fixed effects falls precipitously.

Joint work with Francesco Del Prato; VisitINPS 2023 project.

Abstract. In local labor markets, workers often move at early stages of their careers from lower-paying firms that provide them training, to better-paying, specialized firms. We call this mechanism “human capital value chain” and we document its implications on both workers’ wage paths and local agglomeration externalities.

Extended abstract (for conference submissions) available on request.

Abstract. I consider the problem of estimating the parameters of a game where players are allowed to play correlated equilibria (Aumann, 1974). I show that the existence of correlation between strategies is testable, and I develop an empirical application of the proposed estimator to assess spatial collusion in airline entry.