Variable Differencing and GMM Estimation with Panel Data with Errors-In-Variables

Procedures for estimating a linear single-equation model by means of panel data with errors-in-variables are considered. To eliminate fixed individual heterogeneity, the equation is differenced across one or more than one periods. The differenced equations can be estimated by using as instruments for the differenced error-ridden regressor, level values of this regressor (both leads and lags). General Method of Moments (GMM) procedures for the complete system of differenced equations are considered.