This paper investigated the performance of some Instrumental Variable (IV) estimators and Generalized Method of Moment (GMM) estimators of a dynamic panel data model with random individual effect. The bias and root mean square error criteria were used to assess the sensitivity of the estimators for a serially correlated error term.Monte Carlo study was performed to study the impact of sample size on the performance of different estimators using four different generating schemes for the serial correlation of the error term, namely autoregressive of order one (AR(1)) , autoregressive of order two (AR(2)), moving average of order one (MA(1)) and moving average of order two (MA(2)). The results of the simulation showed that AndersonHsiao Instrumental Variable Estimator in difference form( AH(d)) performed better when the time dimension is small while the one step Arellano-Bond Generalized Method of Moment (ABGMM(1)) outperformed other estimators when the time dimension is large. The biases of most of the estimators improve as the time dimension increases except in some cases. The effect of serial correlation is minimal using different generating procedures.
Digital Object Identifier (DOI)
T. Olajide, Johnson and E. Olubusoye, Olusanya
"Estimating Dynamic Panel Data Models with Random Individual Effect: Instrumental Variable and GMM Approach,"
Journal of Statistics Applications & Probability: Vol. 5
, Article 7.
Available at: https://dc.naturalspublishing.com/jsap/vol5/iss1/7