s-news
[Top] [All Lists]

Re: Johansen Cointegration in S-plus

To: "Ellerton, Tim" <Tim.Ellerton@morganstanley.com>, "s-news@lists.biostat.wustl.edu" <s-news@lists.biostat.wustl.edu>
Subject: Re: Johansen Cointegration in S-plus
From: Guy Yollin <guy.yollin@rotellacapital.com>
Date: Thu, 25 Jun 2009 08:59:25 -0500
Accept-language: en-US
Acceptlanguage: en-US
In-reply-to: <3C171B9F38D32E458D7EFAE90D77A27A25A117FB57@LNWEXMBX0141.msad.ms.com>
References: <3C171B9F38D32E458D7EFAE90D77A27A25A117FB57@LNWEXMBX0141.msad.ms.com>
Thread-index: Acn05d2aT7zX8sf+SBGbQ8KvH+RezQAtKqEQ
Thread-topic: Johansen Cointegration in S-plus

Tim,

 

S-Plus only has tables from Osterwald-Lenum (1992) of the 95% and 99% critical  values for the various cases, it does not compute p-values.

 

You can find a number of places on the web that have the Osterwald-Lenum critical values for 50%, 80%, 90%, 95%, 97.5%, and 99%; for example:

 

http://www2.warwick.ac.uk/fac/soc/economics/staff/faculty/jeremysmith/manual/statisticaltables.pdf

 

If you really want to calculate p-values, I would suggest the following paper:

 

MacKinnon, Haug, and Michelis,

"Numerical distribution functions of likelihood ratio tests for cointegration,"

Journal of Applied Econometrics, 14, 1999, 563-577

 

Available here:

 

http://www.econ.queensu.ca/faculty/mackinnon/papers/lrctest.pdf

 

And their FORTRAN software which calculates critical values and p-values:

 

http://www.econ.queensu.ca/faculty/mackinnon/johtest/

 

Note, their results are somewhat different from Osterwald-Lenum.

 

Best,

 

-- G

 

 

 

From: s-news-owner@lists.biostat.wustl.edu [mailto:s-news-owner@lists.biostat.wustl.edu] On Behalf Of Ellerton, Tim
Sent: Wednesday, June 24, 2009 9:07 AM
To: s-news@lists.biostat.wustl.edu
Subject: [S] Johansen Cointegration in S-plus

 

I want to use the s-plus coint function and get p.values for the trace/max eig. tests.  Is this possible?

 

Many thanks,

 

Tim

 

coint.johansen <- function(x, criterion="BIC", trend="rc", level=0.05, na.rm=T){
 n <- nrow(x)
 # Order Selection
 ord.choice <- VAR(y, max.ar=6, criterion=criterion)
 #cointegration test
 coint.tst <- coint(y, trend=trend, lags=max(1,ord.choice$ar.order-1)
 summary(coint.tst)

 

 #trace test and max eig. tests - Can I get the p-values here???

 

 #vecm
 vecm.fit <- VECM(cointst.rc, coint.rank=1)
 summary(vecm.fit)   


NOTICE: If received in error, please destroy, and notify sender. Sender does not intend to waive confidentiality or privilege. Use of this email is prohibited when received in error. We may monitor and store emails to the extent permitted by applicable law.

<Prev in Thread] Current Thread [Next in Thread>