Analyst Earnings Revisions as Signals for Chinese Industry Rotation
Summary
This report examines whether analyst expectations data can help rank industries for rotation. It groups signals into analyst coverage and forecast disagreement, the direction of earnings revisions, and the size of forecast changes. Coverage and disagreement show little standalone value in the analysis; the report notes that disagreement may be more informative when paired with forecast levels or revision direction. Downward revisions appear more useful than upward revisions, with downgrade proportion and downgrade breadth highlighted as candidate signals. Earnings adjustment measures perform better than valuation adjustment measures, and measures spanning more than one month outperform shorter windows, though revision direction is stronger overall.
A composite of downgrade proportion and breadth is reported to produce a 17% annualized long-short return spread and a 67% monthly win rate. Adding the sentiment signal to an existing industry-rotation model is associated with a 12.07% correlation with next-month industry returns, a 74% monthly win rate for the long-short spread, and 13.8% annualized long-only excess return. The report also gives a 2.3% maximum drawdown relative to an equal-weight industry portfolio. These are historical backtest findings from a specific data source and model; the document warns that patterns and factors may stop working.
Key ideas
- Analyst expectation signals are grouped into coverage and disagreement, revision direction, and revision magnitude.
- Coverage and forecast disagreement show limited standalone usefulness in the reported analysis.
- Downward revision proportion and breadth are identified as stronger industry signals than upward revisions.
- Revision direction performs better overall than revision magnitude, while longer adjustment windows outperform short ones.
- Combining downgrade measures with an existing rotation model improves the reported historical results, but factor effectiveness may change.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.