Am 01.02.2019 um 04:15 schrieb Allin Cottrell:
So I think we await a fully-specified example of disagreement between
R::urca and R::tsDyn or gretl.
I have now used Reynaldo's original example and turned it into a script.
The data are attached in gretl format for convenience.
(Note that I had difficulties getting the exogenous variable into
tsDyn's rank.test, I believe there may be a small bug there. Hence my
workaround with as.matrix and drop=F.)
coint2 6 GDP NDB; Bdum --quiet # Bdum dummy renamed from bnprivfinance
foreign language=R --send-data
cidat <- gretldata[, c("GDP","NDB")]
exo <- as.matrix(gretldata[, c("Bdum")], drop=F) # bnprivfinance
ve <- VECM(cidat, lag=5, estim="ML", include="const",
ve_test0 <- rank.test(ve, r_null=0, type="trace")
ve_test1 <- rank.test(ve, r_null=1, type="trace")
test2 <- ca.jo(cidat, type="trace", K=6, ecdet=c("none"),
Here's (parts of) the output:
Gretl (German, but I guess it's obvious):
Rang Eigenwert Trace-Test p-Wert Lmax-Test p-Wert
0 0.30870 26.346 [0.0006] 20.674 [0.0033]
1 0.096321 5.6718 [0.0172] 5.6718 [0.0172]
Test of rank 0 versus 2 (trace test), p-value: 0.0005827881.
Test of rank 1 versus 2 (trace test), p-value: 0.01724008.
test 10pct 5pct 1pct
r <= 1 | 5.67 6.50 8.18 11.65
r = 0 | 26.35 15.66 17.95 23.52
So that's the differing conclusions from the OP, and Gretl and tsDyn
again agree, urca doesn't. (And BTW, I'm speculating that a comparison
of 26.35 with urca's 1% crit. value 23.5 wouldn't give a p-value below
.001, but who knows).
Finally, let me add that this is just a comparison of the mechanical
workings of the tools. All those results are likely invalid because the
test is not valid in general with such an unrestricted exogenous
variable which is not an impulse dummy, but is modeling some breaks or