Unwinding Dependent Variable Transformations

[jaylamb20]

Hello,

When forecasting from an equation object with a dependent variable specified by expression, EViews gives users the option to forecast the base series or the expression.

For example in "ls d(rx_aud) c ar(1)", EViews will give the option to produce a forecast of RX_AUD or of the differenced series.

Is there are function or proc in EViews for mapping between these expressions and their most basic forms? E.g. I'd like to be able to take predicted values of an expression like "log(rx_aud)-@lag(log(rx_aud),4)" (coming out of a pool estimation) and map them back to predicted values of RX_AUD.

I can hard-code the unwinding of expressions for any specific case, but I'm hoping to write something more general.

Thank you for your help,

-James

[EViews Gareth]

Nope, there's nothing, you'll have to unwind it manually :/

[jaylamb20]

I thought as much but figured it couldn't hurt to ask. Thanks Gareth.

Cheers,

-James

[EViews Gareth]

The internal works that figure out the inverse transformations are black magic. It would be unwise of us to unleash them on the public.

[jaylamb20]

Ha of course, totally understandable. Keep your magic to yourself and I will invent my own.