Hi,
I´m a brazilian student.
The purpose of my research is using AIDS model to estimate economies of scale in brazilian banks.
In doing so it´s necessary to adapt AIDS model including other regressors, not only prices of inputs. At the same time it is included one equation for capital demand.
So I have a system of four equations. According to my fidings, the best way to do that is using nlsur. Neverthless I don´t have a cross sectional dataset. My data is an unbalaced
panel (with gaps) corresponding 66 financial conglomerates with quartely observations from 2001 to 2009. Since I have 90 parameters in each equation is not possible to use cross sectional
analysis.
In this sense, I think that is not correct pooling all thee observations and then use nlsur like a pooled cross section.
The new information is that I found the command xtsur, neverthless I read that, despite the fact that this is ok for unbalanced panels, it uses an appoach of RE. Neverthless, apllying
the Hausman Test to the isolated equations, FE is more appropriate.
Someone could help me?
Thanks for the attention
Antonio Neto
I´m a brazilian student.
The purpose of my research is using AIDS model to estimate economies of scale in brazilian banks.
In doing so it´s necessary to adapt AIDS model including other regressors, not only prices of inputs. At the same time it is included one equation for capital demand.
So I have a system of four equations. According to my fidings, the best way to do that is using nlsur. Neverthless I don´t have a cross sectional dataset. My data is an unbalaced
panel (with gaps) corresponding 66 financial conglomerates with quartely observations from 2001 to 2009. Since I have 90 parameters in each equation is not possible to use cross sectional
analysis.
In this sense, I think that is not correct pooling all thee observations and then use nlsur like a pooled cross section.
The new information is that I found the command xtsur, neverthless I read that, despite the fact that this is ok for unbalanced panels, it uses an appoach of RE. Neverthless, apllying
the Hausman Test to the isolated equations, FE is more appropriate.
Someone could help me?
Thanks for the attention
Antonio Neto