The reduced form of the local level model with conditionally heteroscedastic GARCH(1,1) noises is analyzed. We show that the IMA-GARCH model is a good alternative but its conditional heteroscedasticity is weaker than this of the unobserved disturbances.
展开▼