Interdependent Series

[tranquyl]

Hi

I am trying to create series which depend on each other, These are the 2 types of series blocks I'm trying to generate.

Example 1:
Z(0) = initial Z = 0.1
Y= alpha - 2*Z
Z = Z(-1)+Y(-1)

When I code the above using frml, an error message "circular reference" is displayed. Given the initial value, the interdependent series should be able to solve. We are trying to avoid using "model" as we have many sets of variables that are interdependent.

Example 2:
For series W, it has one formula for smpl 1980-1990 and another formula from 1990-2015. Hence, I know frml cannot be used. But how could I code this please?

smpl 1990 2015
W = V/E
We have data for V and E series from 1990-2015. So we can solve for W.

smpl 1980 1989
W = W(+1) + 2*S - 0.5*F where S and F are series.
By right, this should be solvable and since we have W(1990) from the previous formula. However, when i try this, I get NA for the entire W series.


We are trying to shift our Excel spreadsheets into EVIEWS and would appreciate any advise on how we can convert the series described above.

Thank you for any help!

[EViews Gareth]

The first cannot be done without a model.

The second requires a genr(r) command:

Code: Select all

create a 1980 2015 smpl 1990 2015 series e=nrnd series v=nrnd series w=v/e smpl 1980 1989 series s=nrnd series f=nrnd genr(r) w = w(1) + 2*s - 0.5*f

[tranquyl]

Dear Gareth

Thank you very much for your help! We understand Example 1 has to be done in model. And the genr(r) command works though we didn’t find the details in the manual.
We are trying to combine Example 1 and 2 as follows and was wondering if you could please advise as the codes we used provides NA values.

The idea is that we would like to generate data that depend on each other but we due to data limitations, we need to extrapolate the data. For example,
smpl 1990 2015
Y= alpha - 2*Z
Smpl 1980 1989
Y = Y(1) + 2*s where s is a known series
Z = Z(-1)+Y(-1)

We tried using @iff but the model solves NA. The codes are
Model mtrial
mtrial.append Y = @iif( @year>1990 , alpha – 2*Z , Y(1) + 2*s)
mtrial.append Z = Z(-1) + Y(-1)
mtrial.solve(f)

Lastly, how do we set initial values in a model please? Would really appreciate your advice please!

[EViews Gareth]

I'm not sure it is possible to do what you want.