# Lag Length from a VAR and Vector Error Correction Model (VECM)

A colleague wrote a paper and I am reviewing it for him to make sure it is good. In the paper, the author estimated a VAR to determine the optimal lag length based on the Schwartz Criterion. Then later in the paper the author estimated a Vector Error Correction Model for men and women. For men, the error correction terms had two lags while for women it is three lags for the error correction terms.

I suspect that the lags from the VAR would be independent of the lags of the error correction terms and the lag of 3 for the women specification provided a better fit of the VECM. In other words, the lag lengths from the VAR and VECM would be independent of each other.

Thoughts?

• While we can deduce what's the question you meant to ask, it's still preferable to write out the question in an objective fashion. – Firebug Sep 21 '17 at 17:25

Finally, the error correction term is included in each equation of the VECM only once. It is either lagged by one or by $p$ where $p$ is the lag order of the VECM; the corresponding representations of the VECM are known as long-run and transitory; it is still the same model, just different representations -- pick the one you like. (I think it is possible to have the error correction term lagged by any $m$ between 1 and $p$, but this is not common practice unless $m$ equals 1 or $p$.) Meanwhile, you should not have the error correction term in multiple lags as you will face the problem of perfect multicollinearity.