st: multivariate cointegration test

[Merijn Groenenboom <m.groenenboom@students.uu.nl>]

Dear Statalisters,

I am writing a research about the effect of economic indicators on stock prices.
First I tested for unit roots. Only some variables were stationary at
level, but all variables were stationary at first differences.
Then I will test for multivariate cointegration using Johansen's
method. But before that I use AIC to determine the appropriate lag
lengths.
However I am not sure I have to use level or first differences in AIC
test and multivariate cointegration test.
I know after these tests I have to use first differences for granger
causality test and impulse response function.
But I think AIC and cointegration test should be done at level.
However I am not 100% sure and to avoid spurious results I hope
somebody with experience can tell me the answer!


Thank you.

Merijn G
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/