Dear group,
I'm a newbie to Stata so please allow me to ask several things concerning my steps when running regression
Usually, within 1 country (let say Country A), i know how to run OLS, FEM, REM, GLS, and then GMM to examine the impact of Monetary Indicators, macro-economic and some bank determinant variables on bank performance (ROA-dependent variable), in which the dataset comprises 25 banks with 10 years timeframe (250 observations). I can easily check all the Collinearity, autocorrelation, and heteroskedasticity before running the regression to find the most suitable method of either OLS or REM or FEM or GLS or GMM.
However, my recent challenge arises when I combine the dataset of another country (Country B), which contains 30 banks and spans a 10-year timeframe, with the dataset of Country A (a total of 55 banks, 550 observations) to conduct an overall regression analysis. My study objective.is to examine the impact of monetary policy on bank performance in country A and Country B, and provide a comparative analysis. So, when doing the regression, i command" bysort country : reg.............." to see and compare the regression results between A and B. But, within this consolidated dataset, how can i test the Collinearity, autocorrelation, and heteroskedasticity among variables for each country? Or i need to create another separate dataset for only Country A and Country B to test? So in sum, i have three stata files? Is it reliable?
My last question is: "am I applying the right approach when using the command "by sort country...." to do the regression and comparison? Or you may have another approach? Could you please suggest the most popular one? I only have two countries, not many.
Please do me a big favour on this struggle.
Thank you very much for your kind help.
Regards
Huy Ngo
I'm a newbie to Stata so please allow me to ask several things concerning my steps when running regression
Usually, within 1 country (let say Country A), i know how to run OLS, FEM, REM, GLS, and then GMM to examine the impact of Monetary Indicators, macro-economic and some bank determinant variables on bank performance (ROA-dependent variable), in which the dataset comprises 25 banks with 10 years timeframe (250 observations). I can easily check all the Collinearity, autocorrelation, and heteroskedasticity before running the regression to find the most suitable method of either OLS or REM or FEM or GLS or GMM.
However, my recent challenge arises when I combine the dataset of another country (Country B), which contains 30 banks and spans a 10-year timeframe, with the dataset of Country A (a total of 55 banks, 550 observations) to conduct an overall regression analysis. My study objective.is to examine the impact of monetary policy on bank performance in country A and Country B, and provide a comparative analysis. So, when doing the regression, i command" bysort country : reg.............." to see and compare the regression results between A and B. But, within this consolidated dataset, how can i test the Collinearity, autocorrelation, and heteroskedasticity among variables for each country? Or i need to create another separate dataset for only Country A and Country B to test? So in sum, i have three stata files? Is it reliable?
My last question is: "am I applying the right approach when using the command "by sort country...." to do the regression and comparison? Or you may have another approach? Could you please suggest the most popular one? I only have two countries, not many.
Please do me a big favour on this struggle.
Thank you very much for your kind help.
Regards
Huy Ngo
