panel data

[ghwaheed]

Dear Sir,

I want to estimate the following equation in first difference

Iijt = ao + bDFSjt + bi DBit+ bj DXjt + eijt


where Iijt = is the dependent variable (say, before tax profits/ta) for domestic bank i in
country j at time t; FSjt is the share of foreign banks in country j at time t (i.e. number of
foreign banks divided by the total number of banks); Bit are bank variables for domestic
bank i at time t ; Xjt are country variables for country j at time t. Further, ao is a constant,
and b, bi, and bj, are coefficients, while ijt is an error term. All regressions include country and time-specific fixed effects.
The estimation is by weighted least squares, with the weight being the inverse of the number of domestic banks in a country in a given year to correct for varying numbers of banks across countries. We report heteroscedasticity corrected standard errors.

Can I do this on E views 5? And can you kindly advise by an example how can I organize my dataset? if its possible can you pls advise on a specific manual or guide ? is it better for me to use stata ? Thanx

[EViews Glenn]

It's hard to tell from your notation and description, but I think so. Since you have three dimensions, you'll have to do some of it with expanded variables (using @expand). How many banks and countries do you have?

Should be structured as a panel with country and bank as the cross-section identifiers.