Learning to Detect Recessions from Pseudo-out-of-sample Tests

Name: Learning to Detect Recessions from Pseudo-out-of-sample Tests
Uploaded: 2026-11-01T23:34:25.538008+00:00
Description: An algorithm was developed to detect US recessions in real time by creating hundreds of millions of classifiers from unemployment and vacancy data. This method identifies all 15 historical recessions without false positives by selecting classifiers on an anticipation-precision frontier, leading to a robust, fast signal.

An algorithm was developed to detect US recessions in real time by creating hundreds of millions of classifiers from unemployment and vacancy data. This method identifies all 15 historical recessions without false positives by selecting classifiers on an anticipation-precision frontier, leading to a robust, fast signal.

Open the full episode