Search, Wage Posting, and Wage Dispersion: A Laboratory Approach (Job Market Paper)
Abstract: This paper uses a laboratory experiment to test the influential Burdett and Mortensen (1998) model, which theoretically demonstrates that wage dispersion can arise among identical firms and workers within a labor search framework. The experiment specifically examines the wage-posting games faced by firms as described in the theoretical model. Results from the laboratory sessions support the model's key prediction: wage dispersion emerges under conditions of on-the-job search. Furthermore, the experimentally observed wage distributions closely align with theoretical predictions, providing robust empirical validation for the Burdett-Mortensen framework.
Learning to Evaluate: Boundedly-Rational Agents in the Kiyotaki-Wright Model
Abstract: In this paper, I modify the Kiyotaki-Wright model to evaluate whether bounded rationality can explain the gap between theoretical predictions and lab experiments evidence. I relax the rational expectations assumption and replace it with boundedly-rational agents. By using agent-based modeling with adaptive learning as learning rule, the agents in the simulation can reproduce the behavior of human subjects in lab experiments. The simulation results not only fit the short-run experimental data but also converge to a new equilibrium. Since the modification is based on the baseline K-W model, this suggests that papers which based on the K-W model might have a new equilibrium under bounded rationality.
The Wage Premium of Communist Party Membership: Evidence from China, with P. Nikolov and K. Acker, Pacific Economic Review, 25(3), 309 338.
Works in Progress
Learning to Forecast with the DMP Model
Using AI Agents in Behavioral Experiments: Evidence from a Wage-Posting Labor Search Game