Exact and approximate solutions to incidental parameter problems.

Panel data are becoming ever more available in economics and other disciplines. In panel data models, researchers often wish to include agent-specific parameters (often called fixed effects) to control for unobserved heterogeneity. While the fixed-effect approach is attractive, many dynamic or nonlinear fixed-effect models face a so-called incidental parameter problem: standard estimation methods (e.g. least squares, maximum likelihood) fail to consistently estimate the model parameters that are common to all agents. This research seeks to extend recently proposed exact or approximate solutions to incidental parameters problems (it depends on the model whether an exact solution is possible at all) in the following empirically relevant directions: the inclusion of time dummies; non-stationary data settings; higher-order bias corrections targeting at vey short panels; linear models for panel data with spatial and temporal dynamics.

Keywords:Econometrics, Panel data, Incidental parameter problem, Fixed effects, Asymptotic bias, Bias reduction

Disciplines:Applied economics, Economic history, Macroeconomics and monetary economics, Microeconomics, Tourism