Hi all R users,
I'm trying to replicate the same results that are given in a published
article after been granted
the same data that the authors use.
I'm having problems to determine the cointegration rank of my data set using
the Johnasen's trace test.
This trace test is already programmed in the package ur.ca and can be found
in the function
ca.jo().
After I run the ca.jo() function on my data set I get the following output:
Values of teststatistic and critical values of test:
test 10pct 5pct 1pct
r<=10 9.71 7.52 9.24 12.97
r<=9 24.18 17.85 19.96 24.6
r<=8 45.16 32 34.91 41.07
r<=7 74.12 49.65 53.12 60.16
r<=6 113.96 71.86 76.07 84.45
r<=5 158.58 97.18 102.14 111.01
r<=4 207.95 126.58 131.7 143.09
r<=3 278.29 159.48 165.58 177.2
r<=2 361.34 196.37 202.92 215.74
r<=1 481.09 236.54 244.15 257.68
r=0 626.23 282.45 291.4 307.64
>From this output I would conclude that the cointegration rank is either 9 or
10.
If I run the same test in Stata I get completly different numbers,
especially for my test values. These values
actually correspond to the values the authors get in the article. From that
output I would conclude
that the conintegration rank is either 6 or 7, which would give me completly
different result
when estimating the VECM model then if I use cointegration rank equal to 9
or 10.
Can anyone explain why this difference occurs?
Thanks in advance.
Best regards,
Emil
--
View this message in context:
http://r.789695.n4.nabble.com/Trace-values-in-the-function-ca-jo-tp4640876.html
Sent from the R help mailing list archive at Nabble.com.
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
