using @cumsum(x(-1) in a regression/forecast model for x

[LAO]

Hello,

I am modeling a variable whose value should depend on a function of its lagged cumulative sum.
Does regressing x on c @cumsum(x(-1)) behave the way I want it to, that is, are predicted values of x summed up, when forecasting?

-James

[startz]

No, the actual values of x, added up, are used for forecasting.

[LAO]

I was worried that would be the case. Do you have any ideas on how I can add up and use the predicted values as the forecast rolls ahead?

[startz]

I missed something. You said that the left hand side variable is x. In that case a dynamic forecast may in fact do what you want. You might want to try out a simple example to see what happens.

(Either that or hope Gareth or Glenn intervene )

[EViews Gareth]

Startz's first answer is correct - the actual values will be added up, I believe. I don't think the forecast engine knows what to do with @cumsum.

The only work around is to create a model and solve the model over the forecast period. The model would look something like this:

Code: Select all

y=@cumsum(x(-1)) x = c(1)+c(2)*y

where Y is a temporary series you've created in the workfile, and c(1) and c(2) are the coefficients from your equation (you can just type the values in manually)

[LAO]

Thank you.