Exercise

Horse race

You are now asked to run a horse race in terms of forecasting accuracy between the two approaches for making rolling GARCH model predictions:

  • garchroll: AR(1) standard GARCH model and student \(t\) distribution
  • gjrgarchroll: AR(1) GJR GARCH model and skewed student \(t\) distribution.

The rolling estimations are implemented using n.start = 2500, refit.window = "moving", refit.every = 500.

The resulting ugarchroll objects are available in the console.

Instructions 1/4

undefined XP
    1
    2
    3
    4
  • Set garchpreds to the table with out of sample predictions in garchroll and print the first three lines.