Source: Fama & French (2006) · Journal of Financial Economics · DOI: 10.1016/j.jfineco.2005.09.009
TL;DR
From the valuation identity, expected returns relate to three variables: book-to-market, expected profitability, and expected investment. Holding B/M and profitability fixed, higher investment implies lower expected returns; holding B/M and investment fixed, higher profitability implies higher expected returns (sample 1963–2003). This groundwork motivated the RMW and CMA factors of the five-factor model.
What anomaly it documents
Predictor: expected profitability (positive) and expected investment/asset growth (negative), given book-to-market.
Direction: returns rise with profitability, fall with investment.
Shape: monotone in each given the others; predicted high-minus-low spreads ~0.11–0.25%/month.
OSAP predictor: OperProf.
How to construct it
Sorting variable: operating profitability and asset growth (investment).
Universe: NYSE/AMEX/Nasdaq common stocks (1963–2003).
Portfolio formation: sorts (controlling for B/M) on profitability and on investment.
Long / short: long profitable / low-investment, short unprofitable / high-investment.
Weighting: value-weighted.
Rebalancing: annual.
Evidence and replication
Period
Notes
Source
1963–2003
profitability positive, investment negative, given value
this paper
OOS (post-2006)
formalized as RMW and CMA in FF5 (2015)
Fama-French 2015
OSAP (OperProf)
positive, replicates
Chen & Zimmermann 2022
Why it might work
Valuation identity: for a given price-to-book and discount rate, higher profitability and lower investment must imply higher expected returns.
Quality / q-theory: profitable, conservative-investing firms behave like a quality tilt.
Limitations and risks
Joint controls needed: clearest when value and the other variable are held fixed.
