If anyone could please suggest...
Posted: Tue May 22, 2012 4:37 am
Dear All,
I know this is not the place for asking regarding Econometric questions, but believe me, for the past 5 months I have been dwelling inside research papers, books, articles and manuals to solve and publish my paper. I am doing Gravity Model of Trade on Pakistan. The workfile is attached herewith. My problem is, no matter which data I take, from WorldBank/IMF/National Sources/UN, the data has multicollinearity and as a result the coefficients and significance levels are wrong. Even when I try to estimate "Trade c gdp distance" and then check for multicollinearity by estimating again using "gdp c trade distance", the R Square is greater than the first estimate. I have used several data sources for GDP, has deflated it, but still the same problem. What is wrong with my data or my method, I am completely blank now. Can anyone of you figure something out which I may have skipped or missed?
Secondly, I read alot about Fixed and Random Effects, in fixed effects we cannot estimate the time invarying or the variables which do not change or time. The method is to run another estimation with dependent variable as the Individual Effects and independent being the time invarying variables. But my question is, where are the individual effects in my Eviews? Is it the resid series? I am using Eviews 7.1
I really need this. Any information, suggestion, comment will greatly help. Thanks alot.
Yasir
I know this is not the place for asking regarding Econometric questions, but believe me, for the past 5 months I have been dwelling inside research papers, books, articles and manuals to solve and publish my paper. I am doing Gravity Model of Trade on Pakistan. The workfile is attached herewith. My problem is, no matter which data I take, from WorldBank/IMF/National Sources/UN, the data has multicollinearity and as a result the coefficients and significance levels are wrong. Even when I try to estimate "Trade c gdp distance" and then check for multicollinearity by estimating again using "gdp c trade distance", the R Square is greater than the first estimate. I have used several data sources for GDP, has deflated it, but still the same problem. What is wrong with my data or my method, I am completely blank now. Can anyone of you figure something out which I may have skipped or missed?
Secondly, I read alot about Fixed and Random Effects, in fixed effects we cannot estimate the time invarying or the variables which do not change or time. The method is to run another estimation with dependent variable as the Individual Effects and independent being the time invarying variables. But my question is, where are the individual effects in my Eviews? Is it the resid series? I am using Eviews 7.1
I really need this. Any information, suggestion, comment will greatly help. Thanks alot.
Yasir