Hi, I am applying GMM (Arellano-Bond estimator) to estimate a dynamic panel in TSP. However, according to TSP, many of my instruments are linearly dependent and my data are singular. Any suggestions to overcome the problem?
