The reading list and the three rules
A factor zoo with 316 species. Maybe ten still feed the room.
Harvey Liu Zhu 2016 catalogued 316 named factors in the academic literature. Most fail out of sample. A few survive because the economic story behind them is real. Reading the survivors apart from the rest is the only edge the retail investor has left.
Post publication decay
McLean Pontiff 2016 JF . 97 anomalies . median decay ~58 pct
The factor zoo
Harvey Liu Zhu 2016 RFS . 316 published factors
Convex AUM flows
Sirri Tufano 1998 JF . ~5x inflow asymmetry winners vs losers
01
Read the publication date alongside the backtest dates. In sample is the story. Out of sample is the verdict. If the backtest ends the year the paper landed, you are reading a researcher's pitch deck. If the backtest extends ten years past the publication and the alpha is still there, the factor has earned the right to be in your portfolio at a smaller size than the brochure recommends.
02
Crowding is a tax that grows with AUM. The same factor in Rs 500 cr of AUM is not the same factor in Rs 50000 cr. Sirri and Tufano 1998 showed that flows into winning funds are roughly five times stronger than outflows from losing funds. The convexity is the crowding mechanic. Read the AUM of the largest ETF tracking the factor before you read its trailing return.
03
If you cannot say why it should persist, expect it to decay further. A factor survives when there is a credible economic story for the premium . a risk that scares real money away, a behavioural bias that does not get arbitraged in a decade, or a friction like illiquidity or short sale constraints that caps the smart money's appetite. No story, no survival. The factor zoo is full of dressed up data mines and they all die quietly on the live track record.
The line
Every factor that worked was first an anomaly. Every anomaly published is a factor in decay. The brochure prints the in sample. The fund pays the out of sample. The gap between them is what the market quietly charges you for reading a famous paper. Size the position to the survivor not to the headline. Read the publication date the way you would read the expiry on a carton of milk.
A backtest is a story. The live track record is the verdict. Read both.
rupeecase.com / learn