Title: Recursive estimation in large panel data models: Theory and practice
Authors: Yanrong Yang - The Australian National University (Australia) [presenting]
Abstract: An iterative least-squares estimation method was previously for large panel data models with unobservable interactive fixed effects. The asymptotic distribution of the iterative estimators was provided under the situation of convergence. However, the impact of iteration on the asymptotic properties of the iterative least-squares estimators was not shown. We show that under the traditional assumptions of fixed interactive effects, the recursive procedure will not necessarily yield consistent estimation if the initial estimator is inconsistent. We then analyze sufficient and necessary conditions for the convergence and divergence of the iterative estimators respectively. Simulation results illustrate various examples involving both convergent and divergent iterative estimators. Empirical applications on OECD and divorce rates in United States are employed.