SERVING THE QUANTITATIVE FINANCE COMMUNITY

 
User avatar
nhawrylyshyn
Topic Author
Posts: 1
Joined: April 30th, 2010, 9:10 pm

Johansen Test for Cointegration - Determining Lags

June 27th, 2013, 8:25 pm

Hi,I'm using Johansen's test for cointegration on intraday data in Matlab. My understanding of this test is that the coefficients are quite sensitive to the lag of the VAR/VECM (Vector Autoregressive / Error-Correction) model you choose. I glanced at a few papers and the web and it was suggested that I fit a VAR model first and then minimize the AIC or SC criteria to select the 'optimal' lag. I found that in most cases using a lag of 1 or 2 would suffice and subsequently fit using this lag structure in the Johansen framework.Does this make sense or what is commonly done here when using the Johansen framework for model selection and lag?Many Thanks.
 
User avatar
chocolatemoney
Posts: 322
Joined: October 8th, 2008, 6:50 am

Johansen Test for Cointegration - Determining Lags

July 22nd, 2013, 2:57 pm

ABOUT WILMOTT

PW by JB

Wilmott.com has been "Serving the Quantitative Finance Community" since 2001. Continued...


Twitter LinkedIn Instagram

JOBS BOARD

JOBS BOARD

Looking for a quant job, risk, algo trading,...? Browse jobs here...


GZIP: On