Glossary

R: Testing weak exogeneity in a VECM (urca package)

A property of a variable indicating that it is determined outside the model and is not correlated with the model's error term. Engle, Hendry, and Richard (1983) distinguished three levels: weak exogeneity, strong exogeneity, and super exogeneity.

title: Exogeneity

slug: exogeneity

Definition

Why It Matters

Exogeneity determines whether a variable can be treated as given for the purpose of statistical inference. Weak exogeneity is sufficient for consistent estimation of model parameters; strong exogeneity additionally permits conditional forecasting; super exogeneity ensures that parameter estimates remain stable under policy regime changes — a requirement for Lucas-critique-resistant policy analysis. Confusing conditional inference with causal claims when exogeneity fails leads to biased estimates and flawed policy recommendations.

Example

In a money-demand equation, income may be weakly exogenous if its marginal distribution contains no information about the demand parameters, allowing efficient estimation without modeling income explicitly. However, if a central bank policy change alters the income–interest rate relationship, weak exogeneity may break down. Only super exogeneity guarantees that estimated parameters remain valid under such regime shifts.

```r

library(urca)

vecm <- ca.jo(data, type = "trace", ecdet = "const", K = 2)

summary(vecm) # Check alpha restrictions for weak exogeneity

```

Related Terms

Software Notes

R: Use urca for weak exogeneity tests within VECM; vars for VAR-based Granger-causality tests (a step toward strong exogeneity). Stata: Use vec with constraints to test weak exogeneity; vargranger for Granger-causality tests. Python: Use statsmodels.tsa.vector_ar.vecm for VECM-based exogeneity tests.