EViews.com

Hi there,

I estimated an ARIMA (2,1,3) model and found that AR(1), AR(2) had both significant coefficients however my MA tests were unexpected - both the MA(2) and MA(3) were significant yet the MA(1) was insignificant. Can someone please explain what that actually means? Should I remove one of the MA lags in my estimation?

Thank you very much for your help

Hi ConfusedaboutEViews,

Generally speaking, when selecting an ARIMA model one of the desirables is to choose a parsimonious model, so, on that basis, removing the MA(1) term from the ARIMA(2,1,3) model certainly makes sense and can be justified. One less parameter to estimate means a slightly more efficient model choice.

Whether or not the ARIMA(2,1,3) model provides an adequate representation of the generating mechanism that you're trying to approximate is, however, another story. Without further details - including the methodology that you're employing - it's difficult to tell if your model selection is optimal.

My hunch tells me that dropping the MA(1) will not change your results too much - maybe a little less smoothing. A good way to find out is to simulate different ARIMA processes. It will also depend on the size of the coefficient(s).

Just as a pointer, you may want to keep an eye on the invertibility conditions and cancelling out effects (AR terms -v- MA terms), often referred to as parameter redundancy.

Good luck,
Graeme