When we find seasonality in one of the lags we need to add another independent variable corresponding to that lag. HOWEVER, Schweser mentions that this does not make the model AR(2). If this doesn’t make it AR(2) then what is an AR(2) model and when do we need it?
AR(2) is x(t) = b0 + b1 * x(t-1) + b2 * x(t-2) + variance AR(2) <> AR(1) + seasonality lag You would use AR(2) if you felt that there was a 2 period relation to the current variable value. The AR(1) + seasonality is used to construct a relation between the current and previous values compensated for a seasonal adjustment. Not sure that I’ve explained that well or not…
just remember if that the lag is being added for seasonality purpsoes it doesnt move it from AR1 to AR2. Only if you add the next consecutive time lagged dependendt variable.