Re: [Gretl-users] Discrepancy between R and Gretl in Johansen cointegration
Am 31.01.2019 um 19:13 schrieb Sven Schreiber:
Am 31.01.2019 um 16:07 schrieb Reynaldo Senra:
See also
https://www.rdocumentation.org/packages/tsDyn/versions/0.9-44/topics/rank...
which also uses Doornik's values. Could you compare those two different
tools in R? (There's no unique R way there.)
OK, so gretl and the tsDyn package for R seem to agree about the
p-values if you run this in gretl:
<hansl>
open denmark
coint2 4 LRM LRY --quiet
###
foreign language=R --send-data
require(tsDyn)
cidat <- gretldata[, 1:2] # again LRM LRY
ve <- VECM(cidat, lag=3, estim="ML", include="const") # lags for
diffs!
ve_test0 <- rank.test(ve, r_null=0, type="trace")
ve_test0
ve_test1 <- rank.test(ve, r_null=1, type="trace")
ve_test1
end foreign
</hansl>
So far, so good. However, I checked the critical values in the source of
the urca package for R (ca.jo), and they also appear roughly in line
with the tables in the paper "Numerical Distribution Functions of
Likelihood Ratio Tests for Cointegration" by MacKinnon, Haug, Michelis
(1998, look on the web).
(After figuring out how the tables are structured, that is...)
One table floating around the web which claimed to reproduce
Osterwald-Lenum wasn't really matching those things, though; despite
urca/ca.jo saying they're using these.
So your question still appears to be a good one!
cheers,
sven