A1012
Title: Estimation of heterogeneous panel data models with an application to program evaluation
Authors: Ke Miao - Fudan University (China) [presenting]
Liangjun Su - Tsinghua University (China)
Xun Lu - Hong Kong University of Science and Technology (Hong Kong)
Abstract: Panel data models with two-dimensional unobserved slope heterogeneity and interactive fixed effects are studied. A two-step approach is proposed to estimate the parameters in the model. In the first step, preliminary consistent estimators of the factors and factor loadings via a nuclear-norm-regularization (NNR) are obtained. In the second step, an iterative procedure is proposed to estimate the parameters of interest. The asymptotic properties of the estimators in each stage are established. The proposed model is applied to estimate heterogeneous treatment effects at both individual and aggregate levels. Monte Carlo simulations show that the proposed estimators perform remarkably well in finite samples compared to some existing methods, such as the synthetic control ones. The method is applied to study the effect of economic liberalization on economic growth and find a positive and significant aggregate average treatment effect on the treated (ATT).
