Panel data dynamics and measurement errors: GMM bias, IV validity and model fit - a Monte Carlo study

摘要: An autoregressive fixed effects panel data equation in error-ridden endogenous and exogenous variables, with finite memory of disturbances, latent regressors measurement errors is considered. Finite sample properties GMM estimators are explored by Monte Carlo (MC) simulations. Two kinds compared respect to bias, instrument (IV) validity model fit: differences/IVs levels, levels/IVs differences. We discuss the impact on estimators’ bias other their distributions changes signal-noise variance ratio, length signal noise memory, strength autocorrelation, size IV set, length. Finally, some practical guidelines provided.