seasonal cointegration

[whippoboy2012]

Hello,

My question concerns the usage of mhegy and/or the CATS package (which I do not have).

My understanding is that mhegy is designed to be used with a single time series, rather than multiple time series.
I am interested in looking for seasonal cointegration between several pairs or series and several triples of series.

I see that there has been some work done since the 1990 article that provides the basis for mhegy, much of it by
Franses (e.g. 1998 Journal of Economic Surveys). However, I do not know if it is straightforward to either modify
mhegy to look at multiple series or whether CATS can handle the situation.

To be clear, I am interested in looking for seasonality in monthly data for two or more series and and also to see
if they are cointegrated. Presumably I could proceed by using mhegy on each of the series and then use one of the
available cointegration routines to look for cointegration using either the seasonally adjusted (i.e. after application
of a seasonal unit root if one is found) or unadjusted series.

Thanks in advance.

William Osterberg

[kaemae]

Hey,

my question is similiar:

Are they multivariate seasonal COintegration tests implemented in CATS? Or is any other procedure for RATS available?

I am interested in looking for seasonality in quarterly/monthly data for three series and and also to see if they are cointegrated.
Looking at the HEGY-Tests the results show that the individual series are integrated at the zero AND seasonal frequencies, so the Johansen-test at zero-frequency ist not sufficient.

Any hints?

Thanks in advance for your reply,
Kate

[TomDoan]

CATS does not include tests for cointegration at seasonal frequencies. Johansen, Soren & Schaumburg, Ernst, 1998.
"Likelihood analysis of seasonal cointegration," Journal of Econometrics, vol. 88(2), pages 301-339 extends the ML analysis to seasonal frequencies. One of Soren's students implemented that in RATS (not in a menu-driven way like CATS). I'll see if I can find that.

[kaemae]

Hello TomDoan,
An Algorithm based on Johansen&Schaumburg would be exactly what I am looking for. In their paper DAHL PETERSEN is named with regard to a RATS procedure. But I could not find anything!
Would be great if you could find anything!

Thanks you so much in advance for your efforts!