manipulation of forecast series

[swingline007]

Hello,

I'm trying to figure out an efficient way to manipulate the forecast series of multiple variables. For example I'd like to increase/decrease the out of sample values of the variable apples by 10% to do some "sensitivity testing".

The only way I have come up with is the following:

where there is already a series called apples where the sample period 1990q1 - 2013q1 are my in sample actuals and the forecast period (forecasts already in the series) is 2013q2 2016q4

smpl 1990q1 2013q1
apples_inc10 = apples
apples_dcr10 = apples

smpl 2013q2 2016q4
apples_inc10 = apples*1.1
apples_dcr10 = apples *.90

This give me two series with the same in sample values and with the forecasted values 10% above and and below the orginal forecasted values.

For a high number of variables, is there an easier way to do this? Is there a 'select all' variables command I could use?

Version: 7.2

Thanks,

[EViews Gareth]

Not in EViews 7, no :(

You could always write a program that will do it.

[swingline007]

That's the plan. Though the problem is that the list of variables may change from time to time. So I'm trying to think of an efficient way to write it so there is minimal editing going forward when the variable list changes. Though the only way I can think to write it is just an expanded version of what is posted above. Suggestions?

[swingline007]

Also, does EViews 8 have some sort of "select all variable" command or a feature that I could use that EViews 7 doesn't?

[EViews Gareth]

All versions let you select all variables and put them in a group:

Code: Select all

group mygroup *

You can then just loop through the members of the group.


EViews 8 has an easier way of adjusting the values of a series. You'd still have to loop through the series one by one though.

[swingline007]

I'll give this a try. Thanks.