Johansen Cointegration Test Coefficents vs. OLS Coefficients

[kszynkar]

Hi,

This is probably a silly question, but I have two series that are cointegrated. When I run the cointegration test, the Johansen normalized cointegrating coefficients are different from the coefficients I get using an OLS. Shouldn't these be the same?


Thanks

[kszynkar]

Does my question make sense, or should I provide more details?

[dakila]

For large samples EG will converge to Johansen results. For small sample EG will suffer from endogeneity and serial correlation problem. If you want to use single equation method, it is better to try Stock-Watson or FMOLS.

[kszynkar]

ahhh. Thank you