This website contains preliminary chapters of the book 'Structural Vector Autoregressive Analysis' by Lutz Kilian and Helmut Lütkepohl, Cambridge University Press (to appear in 2017). The copyright rests with the authors. The conmmerical use of these pdf files is prohibited. Click to download.


MATLAB code for selected applications in the book. Click to download. (To be completed).

Chapter 2:

  • LS estimation
  • Restricted GLS estimation [TBA]
  • Bias-corrected LS estimation
  • Lag-augmented LS estimation
  • Table 2.1: Alternative lag-order selection criteria for VAR models

    Chapter 4:

  • Table 4.1: Forecast Error Variance Decomposition for U.S. Real Dividend Growth
  • Table 4.2: Real-Time Risk Measures as of December 2010
  • Figure 4.1: Point estimates of responses of U.S. real dividends to selected structural shocks
  • Figure 4.2: Historical decomposition of the real price of crude oil in percent deviations from the mean
  • Figure 4.3: Contribution of each structural shock to the cumulative change in the real price of oil from January 2003 to June 2008
  • Figure 4.4: Historical counterfactuals for the real price of crude oil from January 2003 to June 2008
  • Figure 4.5: Selected real-time forecast scenarios for the real price of crude oil as of December 2010
  • Figure 4.6: Real-time probability weighted 1-year ahead density forecasts as of December 2010 under different scenarios about the future state of the global economy
  • Figure 4.7: Actual level of global oil production and counterfactual level in the absence of the U.S. shale oil boom
  • Figure 4.8: Counterfactual sequence of flow supply shocks in the absence of the U.S. shale oil boom
  • Figure 4.9: Evolution of the nominal Brent price of crude oil with and without shale oil
  • Figure 4.10: Counterfactual paths of key observables under alternative policy counterfactuals
  • Figure 4.11: Sequence of policy shocks required to implement the KL policy counterfactual

    Chapter 9:

  • Figure 9.1: Responses of the U.S. economy to an unexpected increase in the real price of oil (Cholesky decomposition)
  • Figure 9.1: Responses of the U.S. economy to an unexpected increase in the real price of oil (Iterative solution)
  • Figure 9.1: Responses of the U.S. economy to an unexpected increase in the real price of oil (Rubio-Ramirez, Waggoner and Zha (2010) algorithm)
  • Figure 9.2: Responses of the U.S. economy to an aggregate supply shock in the Keating (1992) model (Iterative solution)

    Chapter 11:

  • Figure 11.1: Responses to technology and non-technology shocks (Cholesky decomposition)
  • Figure 11.1: Responses to technology and non-technology shocks (Rubio-Ramirez, Waggoner and Zha (2010) algorithm)
  • Figure 11.1: Responses to technology and non-technology shocks (Iterative solution)
  • Figure 11.2: Responses to an unexpected U.S. monetary policy tightening (Rubio-Ramirez, Waggoner and Zha (2010) algorithm)
  • Figure 11.2: Responses to an unexpected U.S. monetary policy tightening (Iterative solution)
  • Figure 11.3: Responses to a productivity shock in the baseline model of King et al. (1991) [TBA]

    Chapter 12:

  • Table 12.1: Percentage of Models in Joint Wald Confidence Set Consistent with a Hump-shaped Response Function of Real GNP to an Unexpected Loosening of Monetary Policy
  • Figure 12.1: Example of how resampling overlapping blocks may destroy the dependence structure in the bootstrap data at the point of transition from one block to the next
  • Figure 12.2: Responses of U.S. real GNP to an unexpected loosening of monetary policy: Shotgun plot implied by joint 68% Wald confidence set
  • Figure 12.3: Responses to an unexpected increase in the real price of oil: Shotgun plot implied by joint 68% Wald confidence set with subset of stagflationary responses highlighted
  • Figure 12.4: Global oil market data
  • Figure 12.5: 95% delta method confidence intervals based on bootstrap standard error estimates
  • Figure 12.6: Alternative 95% bootstrap confidence intervals
  • Figure 12.7: Responses of the real price of oil to oil demand and supply shocks with alternative 95% confidence intervals
  • Figure 12.8: 95% delta method confidence intervals based on bootstrap standard error estimates in the overidentified model

    Chapter 13:

  • Numerical example for subrotation algorithm for block-recursive models in Section 13.9.2
  • Numerical example for general algorithm allowing for sign and exclusion restrictions in Section 13.9.2
  • Figure 13.7: Sign-identified oil market model impulse response functions in the modal model and 68% joint HPD regions
  • Figure 13.8: Stuctural impulse responses in the sign-identified oil market model
  • Figure 13.9: Responses to a monetary policy tightening in the original sign-identified model: Response functions in the modal model and 68% joint HPD regions
  • Figure 13.10: Responses to a monetary policy tightening in the modified sign-identified model: Response functions in the modal model and 68% joint HPD regions
  • Figure 13.11: Responses to a monetary policy tightening in the modified sign-identified model