Second Order Differencing

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

bspoon
Posts: 3
Joined: Fri Apr 14, 2017 9:05 am

Second Order Differencing

Postby bspoon » Fri Apr 14, 2017 11:18 am

Hello,

In order to make my data stationary, I had to take the second difference of the natural log of my data. Ex: logy=dlog(y,2).

I used this transformed data to produce a forecast and get values in those terms. I wish to compare my forecasted values to the original values for y in my dataset, so my question is how do I reverse the second order differencing in order to get my forecasted values back into the same terms as my actual values?

Any help would be appreciated.

Thank you!

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13294
Joined: Tue Sep 16, 2008 5:38 pm

Re: Second Order Differencing

Postby EViews Gareth » Fri Apr 14, 2017 11:30 am

Use dlog(y,2) as your dependent variable.
Follow us on Twitter @IHSEViews

bspoon
Posts: 3
Joined: Fri Apr 14, 2017 9:05 am

Re: Second Order Differencing

Postby bspoon » Fri Apr 14, 2017 11:37 am

Yes so the regression that I used for forecasting was with dlog(y,2) as my dependent variable so my forecasted values were in those terms. My original "y" dataset was an index in the 100s so is it possible to transform the forecasted values back in terms of the index in the 100s? I want to compare my forecasted index value with the actual index value if possible.

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13294
Joined: Tue Sep 16, 2008 5:38 pm

Re: Second Order Differencing

Postby EViews Gareth » Fri Apr 14, 2017 1:03 pm

When you use dlog(y,2) as the dependent variable, EViews will forecast y not dlog(y,2)
Follow us on Twitter @IHSEViews

bspoon
Posts: 3
Joined: Fri Apr 14, 2017 9:05 am

Re: Second Order Differencing

Postby bspoon » Fri Apr 14, 2017 1:48 pm

Oh okay I see now. That helps, thank you!

One more thing, when you select the "static" option for a forecast, is that recursive or rolling window? I couldn't find the specification in the Eviews guide.

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13294
Joined: Tue Sep 16, 2008 5:38 pm

Re: Second Order Differencing

Postby EViews Gareth » Fri Apr 14, 2017 5:04 pm

Simple 1-step ahead forecast.
Follow us on Twitter @IHSEViews

Saakshi
Posts: 24
Joined: Fri Jun 10, 2016 10:23 am

Re: Second Order Differencing

Postby Saakshi » Tue May 22, 2018 12:28 am

Hi Gareth,

As you mentioned static forecast means 1 step ahead. What about 4 step ahead forecast?

thanks,
Saakshi

EViews Glenn
EViews Developer
Posts: 2671
Joined: Wed Oct 15, 2008 9:17 am

Re: Second Order Differencing

Postby EViews Glenn » Thu May 31, 2018 11:13 am

Forecast of data at t+4 given information at t.


Return to “Estimation”

Who is online

Users browsing this forum: No registered users and 13 guests