Schweser Exam 3 PM - AR Question #88

In this questions they give 4 lags; Lag 2 and Lag 4 are supposed to be rejected even though the t-stat indicates they are not significant. Can someone explain why? T/G

I think on this one we were supposed to figure out the t stat on each and see if they were significant. I personally didn’t know how to get the std error (now i know 1/sq root # observations (taking one out since there’s one lagged term) which f’d me from the start on figuring out significance. so you do your T tests and none of those lags are significant. i want to say b/c none are significant, you don’t need to add any b/c the model is already correctly specified? bigger picture question- say maybe let’s say lag 2 and 4 were significant- what would you do then? would you add a 2nd lag term if ANY of them were significant? and then do you do same process and test for a 3rd lag term and don’t stop til none are significant? JDV or anyone- they gave you an AR model with a lag one independent variable found significant. guy wonders if he should include 2 lags and lag 4 terms (give you quarterly data). they chart the 4 lags, none of which had significant T stats, and asked about should we use certain ones in the regression. From schweser: If an AR model is correctly specified, the residual terms will not exhibit serial correlation. If the residuals possess some degree of serial correlation, the AR model that produced the residuals is not the best model for the data being studied and the regression results will be problematic. The procedure to test whether an AR time-series model is correctly specified (of the proper order) involves three steps: 1. Estimate the AR model being evaluated using linear regression. For example, if a first-order AR model [i.e., AR(1)] is being evaluated, the following model would be estimated: xt = b0 + b1xt-1 + et 2. Calculate the autocorrelations (correlations) of the model’s residuals. Autocorrelations for the model’s residuals are calculated using lags. For example, for a first-order autocorrelation model, AR(1), with a one-period lag, we test the correlation of residual terms 1 through (T – 1) with residuals 2 through T. For a two-period lag of an AR(1) model, we test the correlation of residuals 1 through (T – 2) with residuals 3 through T. 3. Test whether the autocorrelations are significant. If the model is correctly specified, the autocorrelations of the residuals will not be statistically significant. To test for significance, a t-test is used to test the hypothesis that the residual autocorrelations equal zero.

ps T/G- i suck royalle at quant- any answer i try to help with inevitably is going to come with like 7 more questions since it’s one of my weakest areas overall!