Skip to main content

Cookies on Macro by Mark.We use essential cookies for sign-in, locale, and security. With your permission, we also use optional cookies for preferences, for analytics, and for advertising measurement with the Meta (Facebook) Pixel. Read our cookie policy.

Macro by MarkMacro by Mark
  • Home
Loading…
Loading…
← Models
MacroConceptsPolicyModels

VAR Model

Vector autoregression - a small system where each macro variable is regressed on lags of itself and the other variables in the system.

OverviewGraphMathReferences

Understand the mathematics

Follow the same equations and assumptions one section at a time or in a continuous reading.

Hover or focus a symbol defined in this model to find its exact matches. Select it to keep the highlight; Escape clears the selection.

Reduced-Form VAR Specification

The VAR(p)\text{VAR}(p)VAR(p) model treats nnn macroeconomic variables symmetrically. Each variable is regressed on ppp lags of itself and ppp lags of every other variable in the system. No contemporaneous variables appear on the right-hand side - all contemporaneous interactions are captured by the covariance structure of the innovation vector. The reduced-form representation is the raw statistical object; structural interpretation requires additional assumptions layered on top.

The companion form rewrites the VAR(p)\text{VAR}(p)VAR(p) as a first-order system in a higher-dimensional state space. This representation is the bridge between the coefficient matrices and the stability, forecasting, and IRF properties of the system. All eigenvalues of the companion matrix must lie strictly inside the unit circle for the process to be covariance-stationary.

yt=c+A1yt−1+A2yt−2+⋯+Apyt−p+ut\mathbf{y}_t = \mathbf{c} + \mathbf{A}_1 \mathbf{y}_{t-1} + \mathbf{A}_2 \mathbf{y}_{t-2} + \cdots + \mathbf{A}_p \mathbf{y}_{t-p} + \mathbf{u}_tyt​=c+A1​yt−1​+A2​yt−2​+⋯+Ap​yt−p​+ut​
yt\mathbf{y}_tyt​c\mathbf{c}cut\mathbf{u}_tut​

The VAR(p)\text{VAR}(p)VAR(p) in compact matrix form. Each AiA_iAi​ is n×nn \times nn×n; the full system has n2pn^2 pn2p slope parameters plus nnn intercepts.

(ytyt−1⋮yt−p+1)⏟ξt=(A1A2⋯ApIn0⋯0⋮⋱⋮0⋯In0)⏟Fξt−1+(ut0⋮0)\underbrace{\begin{pmatrix} \mathbf{y}_t \\ \mathbf{y}_{t-1} \\ \vdots \\ \mathbf{y}_{t-p+1} \end{pmatrix}}_{\boldsymbol{\xi}_t} = \underbrace{\begin{pmatrix} \mathbf{A}_1 & \mathbf{A}_2 & \cdots & \mathbf{A}_p \\ \mathbf{I}_n & \mathbf{0} & \cdots & \mathbf{0} \\ \vdots & \ddots & & \vdots \\ \mathbf{0} & \cdots & \mathbf{I}_n & \mathbf{0} \end{pmatrix}}_{\mathbf{F}} \boldsymbol{\xi}_{t-1} + \begin{pmatrix} \mathbf{u}_t \\ \mathbf{0} \\ \vdots \\ \mathbf{0} \end{pmatrix}ξt​​yt​yt−1​⋮yt−p+1​​​​​=F​A1​In​⋮0​A2​0⋱⋯​⋯⋯In​​Ap​0⋮0​​​​ξt−1​+​ut​0⋮0​​
yt\mathbf{y}_tyt​ut\mathbf{u}_tut​

The companion form: a VAR(p)\text{VAR}(p)VAR(p) rewritten as a VAR(1) in an np-dimensional state vector. The eigenvalues of FFF determine stationarity.

∣λi(F)∣<1∀i=1,…,np|\lambda_i(\mathbf{F})| < 1 \quad \forall i = 1, \ldots, np∣λi​(F)∣<1∀i=1,…,np
ppp

Stationarity condition: all eigenvalues of the companion matrix F lie strictly inside the unit circle.

E[yt]=(In−A1−⋯−Ap)−1cE[\mathbf{y}_t] = (\mathbf{I}_n - \mathbf{A}_1 - \cdots - \mathbf{A}_p)^{-1} \mathbf{c}E[yt​]=(In​−A1​−⋯−Ap​)−1c
yt\mathbf{y}_tyt​c\mathbf{c}c

The unconditional mean of the stationary VAR, obtained by setting yt=yt−1=…=E[yt]y_t = y_{t-1} = \ldots = E[y_t]yt​=yt−1​=…=E[yt​] and solving.

Notations

yt\mathbf{y}_tyt​

Endogenous vector

Ai\mathbf{A}_iAi​

Coefficient matrix at lag i

c\mathbf{c}c

Intercept vector

ut\mathbf{u}_tut​

Innovation vector

Σ\boldsymbol{\Sigma}Σ

Innovation covariance

ppp

Lag order

About these symbols

Endogenous vector: n×1n \times 1n×1

Coefficient matrix at lag i: n×nn \times nn×n

Intercept vector: n×1n \times 1n×1

Innovation vector: WN(0,Σ)\text{WN}(\mathbf{0}, \boldsymbol{\Sigma})WN(0,Σ)

Innovation covariance: n×n,PDn \times n, \text{PD}n×n,PD

Lag order:

    Transition checks

    Impact jump
    x0=Bε0,xh=AhBε0(h=1,2,…)x_0=B\varepsilon_0,\qquad x_h=A^hB\varepsilon_0\quad(h=1,2,\ldots)x0​=Bε0​,xh​=AhBε0​(h=1,2,…)

    In a linear state-space representation, B maps the innovation into the contemporaneous state. A delayed variable has a zero row in B and moves only after A propagates the shock. A jump is therefore a timing restriction, not a charting artifact.

    A nonzero point at h=0 means the shock changes the plotted variable on impact. It is not the end of a smooth pre-shock segment.

    Do not smooth through h=0 when the model says the shock enters contemporaneously.

    Decay or hump
    rh=aρhorrh=a he−λhρh,∣ρ∣<1r_h=a\rho^h\quad\text{or}\quad r_h=a\,h e^{-\lambda h}\rho^h,\qquad |\rho|<1rh​=aρhorrh​=ahe−λhρh,∣ρ∣<1

    Stable eigenvalues shrink repeated applications of A. If a variable is reached through another state first, the response rises before the stable decay dominates, producing a hump-shaped path.

    Post-impact movement follows the model recursion. A hump means delayed propagation, not interpolation by the chart.

    Do not treat a hump as evidence of a linear ramp unless the model includes that ramp.

    Variable timing

    Innovation

    ε0≠0,εh=0 for h>0\varepsilon_0\ne 0,\qquad \varepsilon_h=0\ \text{for}\ h>0ε0​=0,εh​=0 for h>0

    The shock is dated at horizon zero. A response at h=0 is an impact effect, not the end of a hidden ramp.

    State vector

    xh=AhBε0x_h=A^hB\varepsilon_0xh​=AhBε0​

    A state can move on impact only through the relevant row of B; otherwise it moves after A propagates the shock.

    Reported response

    rh=CAhBε0+Dεhr_h=CA^hB\varepsilon_0+D\varepsilon_hrh​=CAhBε0​+Dεh​

    The plotted variable can jump directly through D or indirectly through the state vector.

    Keep going

    Read next

    Continue with a related concept, model, or example.

    Agent-based model

    Labor Market ABM

    a contrasting take on VAR

    Contrast it →
    Empirical model

    ARDL / Bounds Testing

    a contrasting take on VAR

    Contrast it →
    Empirical model

    ARIMA

    a contrasting take on VAR

    Contrast it →

    Continue learning

    Concepts

    • Interest rates→
    • Business cycles→
    • Prices and inflation→

    Policy

    • Monetary policy→
    • Inflation targeting→

    Models

    • BVAR→
    • SVAR→
    • VECM→
    Macro by Mark
    Help
    Privacy PolicyTerms of UseEthics and Compliance

    © 2026 Mark Jayson Nation LLC