Hi, I have a dynamic panel data comprising 2400 entities for 36 months? The dependent variable is explained by three exogenous covariates and the lags of the dependent variable. There may be structural breaks in the model as the data belongs to covid term period and beyond. Also there could be seasonality (yearly and possibly quarterly) in the data. I greatly appreciate your inputs into handling each of these things in the model estimation. I have estimated the model first by taking yearly lags into account and then taken rolling means to account for the quarterly seasonality. I ran two models - Arrellano-Bower/Bundell-Bond model for Dynamic panel data analysis and checked for structural breaks separately. After checking for the structural break I introduced a dummy into the first model and I estimated it again. The parameter estimates are very different in both the models and some of the lagged values of the dependent variable are insignificant and the structural break dummy is also not significant. Is there any other way for better estimation of the model?
Thank you,
Sridhar
Thank you,
Sridhar
Comment