4.3 A fundamental representation . . . . . . . . . . . . . . . . . . 264.4 Admissible autocorrelation functions . . . . . . . . . . . . . . 274.5 Multivariate auto- and cross correlations. . . . . . . . . . . . . 30
5 Prediction and Impulse-Response Functions 31
5.1 Predicting ARMA models . . . . . . . . . . . . . . . . . . . . 325.2 State space representation . . . . . . . . . . . . . . . . . . . . 345.2.1 ARMAs in vector AR(1) representation . . . . . . . . 355.2.2 Forecasts from vector AR(1) representation . . . . . . . 355.2.3 VARs in vector AR(1) representation. . . . . . . . . . . 365.3 Impulse-response function . . . . . . . . . . . . . . . . . . . . 375.3.1 Facts about impulse-responses . . . . . . . . . . . . . . 38
6 Stationarity and Wold representation 40
6.1 De
fi
nitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406.2 Conditions for stationary ARMA’s . . . . . . . . . . . . . . . 416.3 Wold Decomposition theorem . . . . . . . . . . . . . . . . . . 436.3.1 What the Wold theorem does not say . . . . . . . . . . 456.4 The Wold MA(
∞
) as another fundamental representation . . . 46
7 VARs: orthogonalization, variance decomposition, Grangercausality 48
7.1 Orthogonalizing VARs . . . . . . . . . . . . . . . . . . . . . . 487.1.1 Ambiguity of impulse-response functions . . . . . . . . 487.1.2 Orthogonal shocks . . . . . . . . . . . . . . . . . . . . 497.1.3 Sims orthogonalization–Specifying
C
(0) . . . . . . . . 507.1.4 Blanchard-Quah orthogonalization—restrictions on
C
(1). 527.2 Variance decompositions . . . . . . . . . . . . . . . . . . . . . 537.3 VAR’s in state space notation . . . . . . . . . . . . . . . . . . 542