Threshold Panel VAR
Threshold Panel Vector Autoregression · Also known as: Panel-VAR with regime switching
The Threshold Panel VAR extends the standard vector autoregression framework to accommodate regime-switching behavior where relationships change when a threshold variable crosses a critical level. Introduced by Hansen (1996) and applied to panels by Caner and Hansen (2001), it allows different dynamic relationships across regimes (e.g., expansions versus recessions) while exploiting the cross-sectional dimension of panel data. This nonlinear framework captures state-dependent policy effects and economic mechanisms.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use threshold-VAR when you suspect structural breaks driven by an observed variable, and you wish to model different dynamic relationships across regimes. It is useful in macroeconomics (studying how policy effectiveness varies with the business cycle) and international finance (examining spillovers in crisis versus normal periods). Ensure the threshold variable is observable and meaningful.
Strengths & limitations
- Captures state-dependent dynamics in a transparent, interpretable framework
- Estimates regime-specific impulse responses revealing differential policy effectiveness
- Panel structure improves efficiency and allows heterogeneous effects
- Regime identification is endogenous, avoiding ad hoc splits
- Estimation is computationally complex (grid search over threshold and VAR parameters)
- Assumes abrupt switching; gradual regime transitions may be misspecified
- Small-sample inference can be unreliable due to threshold non-standard asymptotics
- Limited guidance on threshold variable selection; model can be sensitive to this choice
Frequently asked
How do I choose the threshold variable?
Select a variable with clear economic interpretation as a state indicator. Common choices: unemployment rate (labor market slack), debt-to-GDP (fiscal stress), VIX index (financial stress), or the lagged dependent variable (endogenous switching). Conduct robustness checks across candidate variables.
How do I infer the threshold level itself?
The threshold estimator has non-standard asymptotics (Hansen 1996). Do not use t-statistics for the threshold. Instead, use confidence intervals based on likelihood-ratio inversion or bootstrap methods to quantify uncertainty.
What if I have multiple potential thresholds?
Multi-threshold models are theoretically possible but identification becomes difficult. Start with single-threshold specifications and test whether additional thresholds improve fit significantly before adding complexity.
How do I generate regime-dependent impulse responses?
Estimate impulse responses for each regime separately. Show how a shock propagates through the system in expansion versus recession regimes. Accumulated responses reveal the total dynamic effect over horizons.
Sources
- Hansen, B. E. (1996). Inference when a nuisance parameter is not identified under the null hypothesis. Econometric Theory, 12(3), 386-414. DOI: 10.2307/2171789 ↗
- Caner, M., & Hansen, B. E. (2001). Threshold autoregression with a unit root. Econometric Theory, 17(4), 1-36. DOI: 10.1111/1468-0262.00257 ↗
How to cite this page
ScholarGate. (2026, June 3). Threshold Panel Vector Autoregression. ScholarGate. https://scholargate.app/en/econometrics/threshold-panel-var
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Global VAREconometrics↔ compare
- Panel Smooth Transition RegressionEconometrics↔ compare
- TVP-FAVAREconometrics↔ compare