change of base year - VAR problem

[davidphilipp]

Hi!

I just changed the base year (base1) from 2000:1 to 2005:1 in

{%e}={%e}-@elem({%e},%base1)
{%h1}={%h1}-@elem({%h1},%base1)
{%h2}={%h2}-@elem({%h2},%base1)
{%x1}={%x1}-@elem({%x1},%base1)
{%x2}={%x2}-@elem({%x2},%base1)
{%y1}={%y1}-@elem({%y1},%base1)
{%y2}={%y2}-@elem({%y2},%base1)
series p{%pd}={%x1}-{%x2}

and tried to do the VAR forecast. (It worked with 2000 as base1.) There is a Zero in the base year...

Now I get this error message: "Insufficient number of observations to estimate 20 coefficients per equation in VAR ... "
(I tried it with less coefficients...same error but lower number)

That´s how the VAR looks like:
base=138
smpl @first @first+!base+!j-1
var mvar.ls 1 2 {%dx1} {%de} {%dh1} {%dpd} {%dr1} {%dr2} {%dy1} {%dy2}@ c dpoil dpoil(-1) dpoil(-2)
smpl @first+!base+!j @first+!base+!j+3

mvar.makemodel(m1)
m1.scenario(n,a=_0) "var"
m1.solve(s=d)

I wonder why it does not work, some help would be nice.
Eviews 6

Regards,
David

[EViews Gareth]

Probably best to post your workfile, and the full program.

[davidphilipp]

Some time series ranged from 1970 to 2004 ... that´s why it was not possible.
Thx.