ARIMA Model + Forecast

[rowbro]

Afternoon all,

I am currently struggling with what I suspect is a fairly basic problem, but nevertheless one I can’t seem to crack.

I have built a simple ARIMA model for inflation that estimates inflation components using 7 AR equations. The equations run together using a model, however I am struggling to figure out how to use the model to forecast each equation (3 periods) and then aggregate them (weighted) so as to produce an overall inflation forecast.

Obviously I can do each of these things manually and get the desired result, but it would be great for it all to happen at the click of a button.

Any ideas?

Thanks!

[trubador]

Define the headline inflation as an identity:

Code: Select all

model01.append @identity inf=0.5*inf1+0.3*inf2+0.2*inf3

[rowbro]

Thanks trubador, and please excuse my obtuseness, but how/where do I do this? In the model or in the command line? Also, how do I ensure that the model forecasts each of the equations each time it runs?

Thanks again

[trubador]

If you have already defined and saved your model object into the workfile, then you can do it either way. The model handles equations simultaneously, so you do not have to worry about individual forecasts. Just make sure that you have selected dynamic solve for out-of-sample projections.

[rowbro]

Thanks, I have got that aspect working, but still struggling with the forecasts. The equations I have made don't seem to forecast automatically (I don't know if this is possible), and when the model runs it doesn't forecast each equation. How does one fix this?

Thanks again

[EViews Gareth]

If the equations are in the model, solving the model will forecast from each of them