The cohort ledger and the pipe that drains it
Every factor that worked was first an anomaly. Every anomaly published is a factor in decay.
Five named factors. Window the researcher discovered it in. Window the market traded it in after publication. The third column is what the market quietly took back. The fourth is what survived.
Size . small minus big
Banz 1981 . JFE 9 (1936 to 1980)
Value . high book to market
Fama French 1992 . JF 47 (1963 to 1990)
Momentum . 12 minus 1
Jegadeesh Titman 1993 . JF 48 (1965 to 1989)
Low volatility
Haugen Heins 1975 . JFQA 10 (1926 to 1971)
Profitability . gross profit / assets
Novy Marx 2013 . JFE 108 (1963 to 2010)
The pipe that drains the alpha . four steps from anomaly to commodity
01
Anomaly
A pattern in data nobody is screening for. The alpha is the reward for asking the question first.
>
02
Publication
JF or JFE paper. Citation gives the pattern a name. The name makes it screenable in every shop with a Bloomberg.
>
03
Product
ETF or smart beta fund launches. Retail rupees, pension rupees and institutional rupees can now buy the factor.
>
04
Commodity
AUM funds the arbitrage. Premium compresses to whatever risk or friction stops the next dollar from arriving.
The line the data draws
Across 97 published anomalies the median surviving alpha is roughly 42 pct of the in sample number. McLean and Pontiff 2016 read the same factor in two windows, before and after the paper. The brochure number is almost always the one before. The fund pays the one after. The decay is not a market verdict on the factor. It is the price of admission for everyone who arrived second.