Hello!
In the abdata.dta dataset used in Roodman’s (2009) paper, ‘How to xtabond2’, all variables (n, w, k, ys) are in log form.
From the Stata help manual on gmm post estimation, Stata is able to provide a linear prediction from the model if one specifies the post estimation command -predict-. However, if this is run after xtabond2 in the abdata.dta log-log setting, the prediction is in logs, which is not all that informative.
I would like to know how one uses/manipulates the post estimation command — predict— following xtabond2, or similalry, xtdpdgmm as specified in Kripfganz (2020), when the model is specified as log-log in such a way that one is able to obtain the predictions in unlogged form. My understanding is that one cannot do a simple ‘re-transformation’ to get back to unlogged scale.
Thank you!
Sam
References:
Roodman, D. 2009. How to do xtabond2: An introduction to difference and system GMM in Stata. The Stata Journal. 9(1): 86–136
Kripfganz, S. 2020. Generalized method of moments estimation of linear dynamic panel-data models” 2020 Stata Conference 14, Stata Users Group.
In the abdata.dta dataset used in Roodman’s (2009) paper, ‘How to xtabond2’, all variables (n, w, k, ys) are in log form.
From the Stata help manual on gmm post estimation, Stata is able to provide a linear prediction from the model if one specifies the post estimation command -predict-. However, if this is run after xtabond2 in the abdata.dta log-log setting, the prediction is in logs, which is not all that informative.
I would like to know how one uses/manipulates the post estimation command — predict— following xtabond2, or similalry, xtdpdgmm as specified in Kripfganz (2020), when the model is specified as log-log in such a way that one is able to obtain the predictions in unlogged form. My understanding is that one cannot do a simple ‘re-transformation’ to get back to unlogged scale.
Thank you!
Sam
References:
Roodman, D. 2009. How to do xtabond2: An introduction to difference and system GMM in Stata. The Stata Journal. 9(1): 86–136
Kripfganz, S. 2020. Generalized method of moments estimation of linear dynamic panel-data models” 2020 Stata Conference 14, Stata Users Group.
