I am currently working on a macroeconometric model and one of the goals is to determine if response of variable A to variable C differs from response of Variable B to variable C. How can I address the issue with the use of EViews.
Probably with a combinated study between impulse-response function and granger causality in E-Views.
If you are using time series data on one country, you can try modelling your data using vector autoregression (VAR). From VAR it is easy to extract impulse-response, as our colleague suggested. You can also try structured VAR, where you can put some theoretical restrains on your model. I am sure you will find exact approach for EViews online as it is heavily applied, but make sure data can me analysed in VAR environment.
Are you taking about Cointegration between the variables, in this case you have to use VECM or VAR model as recommended by Tasic.
before you go further in this you have to check the series for stationary, Unit Root using ADF test or PP test ...
log differential between Index and futures in misfit 5.5:
Let LOGI= LOG(Index);
LOGF=LOG(Futures);
SO, the log differential is following:
LDIF=LOGI-LOGF;
and what is the different between first difference and log differential in microfit5.5?
well,
the first difference:
FDI= LOGI-LOGI(-1);
FDF=LOGF-LOGF(-1);
Log differential :
LDIF=LOGI-LOGF;
Additional in Microfit5.5:
ADF for testing stationarity of the model,
AR for testing whether the series Random Walk (has a unit root).
- Badges
- Science topic
- Similar topics
- Mathematics