In order to estimate aVAR, alag length must be used in the estimation. There are many different criteria which can be used to signal the ideal lag length to use.Asteriou & Hall (2011) explain that the following three statistics are mainly used.
The first criterion which assesses the goodness of fit of the model is the Akaike Information criterion, (AIC), developed by Akaike (1974) and is written;
AIC=(RSS/n) e(2k/n)
The Schwarz Criterion (SC) developed by Schwarz (1978) with the following equation;
SC=(RSS/n) e(k/n)
The SCstatistic is likely to penalise complexity within models.
Finally, the Hannan and Quin criterion, (HQC);
HQC=(RSS/n) (ln?n)(2k/n)
The method of selecting the ideal lag length will be the lag number which minimises the statistics of each criterion. Should the tests produce conflicting results, then an explanation into which test that I will use will accompany the results.