[R] Function: VECM (Johansen)
Pfaff, Bernhard
Bernhard.Pfaff at drkw.com
Mon Sep 9 18:40:16 CEST 2002
[message bounced because of "octet-stream" attachment which
are not allowed in our mailing lists;
manually fixed and approved, MM]
Dear R-list,
R: 1.5.1
OS: Windows NT
additional packages needed: tseries
for those of you who are interested, pls. find attached a function for
estimating VECM's by employing the method of Johansen (see for example:
Hamilton, "Time Series Analysis, Chapter 21").
Arguments:
y: data matrix
adf: logical, indicating wether Augmented-Dickey-Fuller Tests
should be calculated or not.
dtt: logical, wether the data matrix y is governed by a
deterministic trend, too.
graph: logical, diagram of fits (level and first differences of the y
series and impulse response functions).
pvar: integer, how many lags the underlying VAR should have. Note,
the number of lags in the VECM is one order smaller than the
VAR-order is.
ira: integer, number of periods for the impulse response
function.
Output:
ALPHA: vector of constants.
PSI0: Error-Correction-Matrix
PSI: Array of matrices for the lagged differences of y.
OMEGA: Variance/Covariance matrix.
ENDOG: Matrix of endogenous variables (1st differences).
FITTEDD: Fitted values in 1st differences.
FITTEDL: Fitted values in levels.
IRA: Array of Impulse Responses
Note, this function runs only in interactive mode: the user has to supply
the number cointegrating vectors after the function returns the trace and
rank statistic.
Note, for Non-Windows users adjust the lines:
if (file.access("graph_y1_L.jpg")==0)
{
system(paste("cmd /c cd ", getwd(), " && del /q *.jpg", sep=""))
}
accordingly to your OS.
Any comments are appreciated.
Bernhard
<<johann.R>>
If you have received this e-mail in error or wish to read our e-mail disclaimer
statement and monitoring policy, please refer to http://www.drkw.com/disc/email/
or contact the sender
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: johann.R
Url: https://stat.ethz.ch/pipermail/r-help/attachments/20020909/98ebb733/johann.pl
More information about the R-help
mailing list