A0834
Title: Least squares estimation of dynamic games with unobserved heterogeneity
Authors: Sorawoot Srisuma - National University of Singapore (Singapore) [presenting]
Abhimanyu Gupta - University of Essex (United Kingdom)
Fabio Sanches - Sao Paolo School of Economics (Brazil)
Abstract: An estimator for dynamic decision models and dynamic games with time-varying unobserved heterogeneity is proposed. The approach is general as the imposition of finite dependence or an instrument is not required. The procedure involves a classification of the conditional choice probabilities via k-means clustering, which is followed by any of the existing two-step estimators in the literature. The focus is on the OLS-type estimators of a prior study for numerical simplicity. The Monte Carlo studies show very good finite sample properties of the estimator.
