Hi,
it's unclear to me whether (and if so, how) it's possible to restrict
simultaneous-equations system sequentially. What I mean is something
like this:
<pseudo-hansl>
toberestricted <- system
...
end system
estimate toberestricted method=ols # unrestricted
restrict
b[1,1] = 0 # for example
end restrict
estimate toberestricted method=sur # one restriction applied
restrict
b[2,1] = 0 # again, an example
end restrict
estimate toberestricted method=sur
</pseudo-hansl>
What I would like is that the last estimation command takes into account
both previously defined restrictions, but that the test just evaluates
the last restriction, relative to the system with the previously imposed
first restriction.
(There is a similar mechanism in the VECM context when restricting the
cointegration relations; in the GUI there is a tickbox for deciding
whether to test all restrictions, or just the last ones.)
Thanks,
sven