Uncertain Short-Run Restrictions
and Statistically Identified
Structural Vector Autoregressions

Welcome to an interactive illustration

This interactive demo invites you to explore a combination of a statistical identification approach with potentially invalid short-run zero restrictions in SVAR models. Our ridge-regularised estimator shrinks towards economically motivated zero restrictions when the data support them and relaxes them when evidence suggests otherwise

With only a handful of sliders and switches you can:

• Generate data under alternative data-generating processes.
• Observe how estimates based on invalid short-run restrictions lead to biased esimates.
• Explore how estimates based on non-Gaussianity tend to be volatile.
• Explore how our combination of statistical identification with short-run restrictions balances efficiency and robustness

Tweak a control and watch plots, matrices, and LaTeX formulas update in real time.