Structural Break Explorer

True Data-Generating Process

Set the true break — or set magnitude to zero for the null of no break — and watch whether the QLR test finds it.

Break Design

Break location (fraction of T) 0.50

Break magnitude (mean shift) 1.50

Noise volatility σ 1.00

Sample

Sample size T 200

15% trimming (Andrews 1993) · k=1 (mean-shift test) · 5% critical value ≈ 7.17

—

Adjust the sliders to run the test.

Simulated Series (true break marked in terracotta, estimated break in charcoal)

Chow F(τ) Statistic Sequence — QLR is the supremum over trimmed candidates