mean group estimator
Posted: Fri Jul 06, 2018 7:47 am
Dear Sir,
I am working with panel data with 11 countries. In order to estimate the MG estimator, I run 11 individual OLS regressions and calculated the unweighted average of the coefficients of each parameter in the model. How do I calculate the Standard Error of the Mean Group estimator? When presenting the MG estimator most authors include the SE as well.
And if I may ask a second question, besides my first method, that of estimating the individual regressions for each country, I also pooled the data and estimated the coefficients using the Pooled Least Squares Method and then manually I calculated the average of the coefficients to arrive at the MG estimator. When doing this, the coefficients are identical to the ones I got when I used the first method (and therefore also the MG) however the standard errors or the coefficients are different. Which SE should I present, the ones resulting from the individual regressions or the ones from the pooled regression?
Thank you
I am working with panel data with 11 countries. In order to estimate the MG estimator, I run 11 individual OLS regressions and calculated the unweighted average of the coefficients of each parameter in the model. How do I calculate the Standard Error of the Mean Group estimator? When presenting the MG estimator most authors include the SE as well.
And if I may ask a second question, besides my first method, that of estimating the individual regressions for each country, I also pooled the data and estimated the coefficients using the Pooled Least Squares Method and then manually I calculated the average of the coefficients to arrive at the MG estimator. When doing this, the coefficients are identical to the ones I got when I used the first method (and therefore also the MG) however the standard errors or the coefficients are different. Which SE should I present, the ones resulting from the individual regressions or the ones from the pooled regression?
Thank you