Page 1 of 1

Re: rolling chow test

Posted: Thu Apr 22, 2010 11:46 am
by EViews Gareth
It isn't particularly clear what you're trying to do with the line:

Code: Select all

smpl 2 !i
Perhaps it is meant to be:

Code: Select all

smpl @first+!i @last

Re: rolling chow test

Posted: Thu Apr 22, 2010 1:41 pm
by marienne
actually im just so lost with what i am dealing with. im trying to make two matrix, one for ssr and one for p values.
matrix(500,6) results
matrix(500,2) pvalues

scalar pval

smpl @all
equation eq1.ls wage c @trend
results(1,1) = eq1.@ssr

from here i have the unrestricted ssr. then i divide the sample into two
for !horizon=10 to 149
smpl 1970m1 1970m1+!horizon
equation eq2.ls wage c @trend
results(!horizon,2) = eq2.@ssr
next

for !horizon=10 to 149
smpl 1982m10 1982m10+!horizon
equation eq3.ls wage c @trend
results(!horizon,3) = eq3.@ssr
next

then i try to make the restricted ssr
results(!horizon,4) = results(!horizon,2) + results(!horizon,3)
results(!horizon,5) = ((results(1,1)-results(!horizon,4))/2)/(results(!horizon,4)/ (nobs-1-4))
results(!horizon,6) = @fdist(results(!horizon,5),2,nobs-1-4)

if results(!horizon,6) < 0.05 then
pvalues(!horizon,2) = 1
else
pvalues(!horizon,2) = 0
endif
else
pvalues(1,1) = 0
but it does not work..totally stuck

Re: rolling chow test

Posted: Thu Apr 22, 2010 1:57 pm
by EViews Gareth
You'll have to be more specific than "it does not work".

Have you read the Programming Guide?