Question

We would like to test the accuracy of a no-change benchmark for output growth relative to a VAR model for the past 40 years. Below is the result of the test of equal predictive accuracy measured in terms of mean squared forecast error. What do you conclude, which model is better? Comment. Here the 'fe_diff' is defined as squared forecast errors implied by the no-change forecast minus the squared forecast errors implied by the VAR. reg = 1m(fe_diff ~ 1) summary(reg) Call: 1m formula = fe_diff ~ 1 Residuals: Min 1Q Median 3Q Max -5.7154 -0.2042 -0.0621 0.1989 5.0282 Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) 0.10293 0.08623 1.194 0.234 Residual standard error: 1.091 on 159 degrees of freedom

          We would like to test the accuracy of a no-change benchmark for output growth relative to a VAR model for the past 40 years. Below is the result of the test of equal predictive accuracy measured in terms of mean squared forecast error. What do you conclude, which model is better? Comment. Here the 'fe_diff' is defined as squared forecast errors implied by the no-change forecast minus the squared forecast errors implied by the VAR.

reg = 1m(fe_diff ~ 1)
summary(reg)

Call: 1m formula = fe_diff ~ 1
Residuals: Min 1Q Median 3Q Max 
-5.7154 -0.2042 -0.0621 0.1989 5.0282

Coefficients: Estimate Std. Error t value Pr(>|t|) 
(Intercept) 0.10293 0.08623 1.194 0.234

Residual standard error: 1.091 on 159 degrees of freedom

we would like to test the accuracy of a no change benchmark for output growth relative to a var model for the past 40 years below is the result of the test of equal predictive accuracy measu 15341

Added by Sarah L.

Question

Please give Ace some feedback