pmpp (version 0.1.0)

ssys_gmm: Suboptimal multi-step System-GMM estimator for AR(1) panel data model

Description

Computes an enhanced version of the Blundell-Bond (System-GMM) estimator for panel data by means of replacing the standard GMM-weighting matrix by its sub-optimal version, thus increasing estimator's efficiency.

Usage

ssys_gmm(Y, model = c("onestep", "twosteps", "threesteps"))

Arguments

Y

matrix of size (T x N) with the dependent variable

model

one of: onestep, twosteps, threesteps; more steps should increase efficiency, but might be computationally infeasible (a singular matrix needs to be inverted); if this is the case, generalised inverse is used

Value

The estimated value of the auto-regressive parameter.

References

Youssef, A. and Abonazel, M. (2015). Alternative GMM estimators for first-order autoregressive panel model: An improving efficiency approach. MPRA Paper No. 68674; Forthcoming in: Communications in Statistics - Simulation and Computation, https://mpra.ub.uni-muenchen.de/68674/1/MPRA_paper_68674.pdf