An Introduction to Factor Investing
A factor is a characteristic shared by a group of securities that has historically been associated with differences in return. Factor investing is the practice of describing portfolios in terms of those characteristics rather than in terms of the individual companies held. The shift sounds academic. Its practical consequence is that two portfolios holding entirely different stocks can be shown to be nearly the same bet, and a manager's apparent skill can often be reproduced by a mechanical rule.
Starting point: the single-factor world
The Capital Asset Pricing Model, developed in the 1960s, held that the expected return on any asset depended on one thing: how much it moved with the overall market, a quantity called beta. Risk specific to an individual company was assumed to be diversifiable and therefore uncompensated. The model was a genuine intellectual achievement and it gave the field its basic vocabulary.
It also did not fit the data especially well. Through the 1970s and 1980s researchers kept finding groups of stocks whose returns the model could not explain. Small companies outperformed what their betas predicted. So did companies trading cheaply relative to book value or earnings. These were called anomalies, a word that quietly assumes the model is right and the world is misbehaving.
The three-factor model
Fama and French (1993) proposed the obvious response: if size and value systematically explain returns, put them in the model. The three-factor model expresses a portfolio's excess return as its exposure to three things rather than one.
Mkt-Rf is the market's return above the risk-free rate. SMB, Small Minus Big, is the return of small-capitalisation stocks minus large ones. HML, High Minus Low, is the return of high book-to-market (value) stocks minus low book-to-market (growth) stocks. The coefficients describe how much of each exposure the portfolio carries, and alpha is whatever is left over once those three are accounted for.
Why this reframing mattered: a fund manager who beat the market by owning small, cheap companies was previously credited with skill. Under the three-factor model, that return is attributable to known exposures anyone can buy cheaply. Alpha became a much smaller and much better defined quantity.
The five-factor extension
In 2015 Fama and French added two more factors. RMW, Robust Minus Weak, captures operating profitability: among companies with similar valuations, more profitable ones have tended to do better. CMA, Conservative Minus Aggressive, captures the investment effect: companies growing their asset base rapidly have tended to subsequently underperform more conservative ones.
Both have a foundation in valuation theory rather than being purely empirical discoveries, which is part of why they were adopted relatively quickly. A notable consequence of the extension is that HML became partly redundant in the five-factor specification, since profitability and investment absorb some of what value was previously capturing.
Momentum, and why it is not in the model
Momentum, the tendency for recent winners to keep winning over intermediate horizons, is among the most robust patterns ever documented, appearing across countries, asset classes, and time periods. Carhart (1997) added it to produce a four-factor model, and many practitioners use it routinely.
Fama and French have nonetheless declined to include it, largely because it resists a convincing risk-based explanation and because its high turnover makes it costly to implement. Its exclusion is a genuine and ongoing disagreement in the field rather than a settled matter.
What these models are good for
The most reliable use of a factor model is diagnostic. Run a portfolio's returns against the factors and you learn what it is actually exposed to, which is frequently not what its description implies. Two funds with different names and holdings often turn out to carry nearly identical exposures. A strategy that appeared to add value often turns out to have been a leveraged bet on a single well-known factor.
The least reliable use is forecasting. These models describe how returns have varied historically. They do not establish that the premiums will persist, and the size factor in particular is a cautionary example of a premium that weakened sharply after it became widely known. The other articles in this section take up each factor in turn, along with the practical problems that appear when the theory meets real trading costs.
Fama, E. F., & French, K. R. (2015). "A five-factor asset pricing model." Journal of Financial Economics, 116(1), 1-22.