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

VECM Model

Vector error-correction model - for cointegrated macro series where long-run equilibrium relationships pull the system back.

OverviewGraphMathReferences

Graph workspace

Cointegrating equilibrium with error correction dynamics

Graph workspace

The defining object of the VECM is the interplay between the long-run equilibrium (the cointegrating relationship) and the short-run adjustment dynamics. A dual-panel display - top panel showing two I(1) series and their equilibrium, bottom panel showing the stationary equilibrium error and adjustment speed - makes both layers visible simultaneously.

Chart axis metadata
xTimequarters · 1-100yLevel / Equilibrium errorindex · -3 to 3 for equilibrium error
A cointegrated pair, drawn on the same
scale static chart preview showing First series, Second series, scaled by beta
Equilibrium error static chart preview showing Equilibrium error, Long-run equilibrium
xTimequarters · 1-100yLevel / Equilibrium errorindex · -3 to 3 for equilibrium error

Current results

Half-life of adjustment

Explain: Half-life of adjustment
h∗≈ln⁡(0.5)ln⁡∣1+α1−βα2∣h^* \approx \frac{\ln(0.5)}{\ln|1 + \alpha_1 - \beta \alpha_2|}h∗≈ln∣1+α1​−βα2​∣ln(0.5)​

Quarters until the equilibrium error decays to half its initial value. Each quarter the error shrinks by the factor 1+α1−βα21 + \alpha_1 - \beta \alpha_21+α1​−βα2​: y1y_1y1​ closes the gap from one side (α1<0)(\alpha_1 < 0)(α1​<0) and y2y_2y2​ from the other (α2>0)(\alpha_2 > 0)(α2​>0), so a positive α2\alpha_2α2​ shortens the half-life. The short-run term γ11\gamma_{11}γ11​ is left out of this approximation, and it is blank when the error does not decay. A half-life of 4-5 quarters is typical for macro error correction models.

3.11

quarters

Equilibrium error std dev

Explain: Equilibrium error std dev
sd(ect)=sd(y1,t−βy2,t)\text{sd}(ec_t) = \text{sd}(y_{1,t} - \beta y_{2,t})sd(ect​)=sd(y1,t​−βy2,t​)

Measures how far the system typically wanders from equilibrium. Larger values indicate weaker or slower error correction.

0.9892

index units

Long-run variance ratio

Explain: Long-run variance ratio
Var(y1,t)Var(ect)\frac{\text{Var}(y_{1,t})}{\text{Var}(ec_t)}Var(ect​)Var(y1,t​)​

Variance of the level of y1y_1y1​ over the variance of the equilibrium error y1−βy2y_1 - \beta y_2y1​−βy2​: how far the shared trend carries y1y_1y1​ compared with how far the pair strays from its long-run relation. Higher means a tighter cointegrating relation.

7.17

ratio

Weak exogeneity test

Explain: Weak exogeneity test
∣α2∣<threshold⇒y2 weakly exogenous|\alpha_2| < \text{threshold} \Rightarrow y_2 \text{ weakly exogenous}∣α2​∣<threshold⇒y2​ weakly exogenous

If α2\alpha_2α2​ is near zero, variable 2 does not respond to the equilibrium error. Inference on β\betaβ can then be conducted in a single-equation framework.

0.0500

coefficient magnitude

Controls

Pulls back quickly
Explain: How fast series 1 closes gaps

Adjustment speed (y₁): -0.15

How quickly the first series responds when the two series drift from their long-run relationship; negative values pull it back.

More negative α1\alpha_1α1​: the equilibrium error in the bottom panel reverts faster, and y1y_1y1​ in the top panel tracks the equilibrium line more closely.

0.05
Explain: Response of series 2 to gaps

How much the second series moves when the two series sit away from their long-run relationship.

Nonzero α2\alpha_2α2​: both series participate in error correction. If α2\alpha_2α2​ has the same sign as α1\alpha_1α1​, both variables chase each other - unusual but possible.

One-for-one
Explain: Size of the long-run link

Cointegrating coefficient: 1

How much the first series moves in the long run for each one-unit move in the second.

Changing β\betaβ rotates the equilibrium line in the top panel and rescales the equilibrium error in the bottom panel.

Changes partly keep going
Explain: Short-run momentum in series 1

Short-run persistence (y₁): 0.2

How much last period's change in the first series carries into this period's change.

Higher γ11\gamma_{11}γ11​: y1y_1y1​ changes are more persistent in the short run, creating smoother trajectories between equilibrium corrections.

0.5
Explain: Size of shocks to series 1

Typical size of the random shocks hitting variable 1; larger values mean more noise.

Larger σ1\sigma_1σ1​: both series become noisier, and the equilibrium error has wider excursions. The cointegrating relationship is harder to see visually.

Keep going

Read next

Continue with a related concept, model, or example.

Theory model

IS-LM Model

a well-connected theoretical model two steps from here

New angle: theoretical model →
Macro by Mark
Help
Privacy PolicyTerms of UseEthics and Compliance

© 2026 Mark Jayson Nation LLC