Question regarding first differences

[Christian_87]

Hello

At the moment I am building an AR(2) modell (times series) were I have a produced a forecast. To achive stationarity I had to take the log first differences of my variables before making the forecast. The forecast has given me forecasted first differences values. My question is how do I recalculate the forecasted values back to its "original values"? I have tried a thing like x1+x1(+1) ( the opposite of first differnces, but it doesent work.)

Thanks in advance!

[startz]

If you make your forecast variable d(x), instead of defining a new variable equal to d(x), EViews will forecast x for you.

[Christian_87]

Thanks for the quick answer!

Iam not shore of what you are meaning exactly. So if I get a forecasted variable that is log diff, should i transform the forecasted variable log diff to its original value? Or is their some kind of function in Eviews that i should use?

[startz]

Try the following regressions and then forecasting from each to see what variables are created by the forecast.

Code: Select all

ls d(x) c z ls dlog(x) c z

[Christian_87]

Thanks!

It works now!