Advanced-Market Overheating Index
Reading a financial cycle
This is the Advanced-Market overheating index from Chen & Svirydzenka's IMF study of financial-cycle early-warning indicators. The goal of this papers is to see if an economy's real equity prices and its output are far enough above their long-run trends to resemble the run-ups that preceded past banking crises.
The paper uses two metrics for 34 advanced economies: equity prices and output. These were the paper's strongest indicators that minimizes missed crises and false alarms, outperforming credit-to-GDP and property prices. The index watches those two indicators and compares them against thresholds calibrated on historical crises, using the Christiano–Fitzgerald filter described in Methodology.
The two signals
Each signal looks at how far the actual level sits above its calculated trend. One fires when real equity prices are more than 6.7% above their calculated trend, the other when real GDP is more than 1.3% above its calculated trend.
| Signal | Measures | Threshold |
|---|---|---|
| Real equity-price gap | real share prices versus their expected values | +6.7% |
| Output gap | real GDP versus its expected value | +1.3% |
The index
Each signal is an on/off flag whether it is above its threshold or not. The index is their weighted sum, so it can only take four values.
The weights (0.497 and 0.503) come from each indicator's crisis-prediction reliability in the paper. They are near-equal by coincidence, so the value is essentially a count of the signals above their thresholds.
Every calculated trend comes from the Christiano–Fitzgerald band-pass filter. The full pipeline, from raw data through where the filter bands come from to the filter mathematics, is laid out under Methodology.