An analyst wants to model quarterly sales data using an autoregressive model. She has found that an AR(1) model with a seasonal lag has significant slope coefficients. She also finds that when the second and third lags are added to the model, all slope coefficients are significant too. Based on this, the best model to use would most likely be an: A) AR(4). B) AR(0). C) AR(1). D) AR(2). The correct answer was A. She has found that all the slope coefficients are significant in the model xt = b0 + b1xt1 + b2xt4 + et. She then finds that all the slope coefficients are significant in the model xt = b0 + b1xt1 + b2xt2 + b3xt3 + b4xt4 + et. Thus, the second model, the AR(4), should be used over the first or any other model that uses a subset of the regressors. -------------------------------------------------------------------------------- Okay fine with the explanation but doesnt schweser say that multi period forecasts are more uncertain than single period forecasts tell me i know but the above is not an AR exactly rght whats the subset of the regressors

multi-period forecasts means mulii-period not multi lags… like we are in the month of june… to get results of sep we need to forecast july, then use july to get aug, then use aug to get sep… this is mulit period…chain rule ar(1) means 1 lag ar(2) means 2 lags this has nothing to do w/ multi period forecasting…

“AR(1) model with a seasonal lag has significant slope coefficients” Is this a problem? I thought we had to try different lags only if the autocorrelation of the residuals had significant coefficients, not the AR(1) model itself. Just when I thought I had figured out this section