0

The question is exactly on the title. I am building a VAR model for a specific commodity. The AIC and BIC suggested 1 lag. However, that coefficient in lag 1 is not significant in my model. Can I still forecast using this model or is it justified to add lags? If I add lags, how much lag needs to be significant until the model is considered a good fit?

kjetil b halvorsen
  • 77,844
Sandy
  • 11
  • 2
  • 1
    https://stats.stackexchange.com/questions/275861/aic-and-bic-criterion-for-model-selection-how-is-it-used-in-this-paper/275948#275948 demonstrates that it can be possible to "translate" AIC to a hypothesis test, and a fairly liberal one, so that such behavior is possible. – Christoph Hanck Nov 15 '22 at 09:05
  • See for instance https://stats.stackexchange.com/questions/523485/why-is-aic-or-bic-commonly-used-in-model-selections-for-time-series-forecasting – kjetil b halvorsen Nov 16 '22 at 01:51

0 Answers0