I am estimating the determinants of Non Performing Loan in Islamic Banking Industry using eviews 8.
I use estimation techniques panel FMOLS and fixed effect estimation techniques to examine the cointegrated variables. Subsequently, when compared to the fixed effect estimation results, there is no constants (C) in panel FMOLS estimation output (while in fixed effect estimation, there is). Would you explain why? Then what distinguishes these two estimation techniques? Which one is better? Thank you so much, your help is much appreciated.
Zira - Did you figure this out? I am not familiar with that software, but by a 'constant,' are you referring to the result of estimating an intercept term? If you expect that as x (regressor or regressors) approaches zero, y should also approach zero, then you should not use an intercept term. In such regression through the origin (RTO), you may also expect heteroscedasticity. - I hope that this relates to your question, but I am not certain. - Cheers! - Jim
- Badges
- Science topic