Is a factor premium a risk or a mistake — and why does it matter?
Two stories explain every factor, and they are not academic. Which one you believe decides whether you expect the premium to keep arriving after you buy it.
The risk story
Cheap companies are cheap because they are in trouble: high debt, falling margins, an industry in retreat. Holding them is holding the risk that the trouble deepens in a recession, exactly when you can least afford it, and the extra return is the pay for bearing that. Small companies are illiquid and fragile in the same way. In this story a factor is a risk premium like the equity premium itself — it persists because the risk persists and somebody has to be paid to hold it. Fama and French have always argued this side.
The consequence is optimistic in the long run and honest about the short: the premium should keep arriving, and it will arrive as pay for years that hurt. A factor that never lost would not be a risk premium.
The behaviour story
Investors extrapolate. They pay too much for companies with exciting stories and too little for boring ones, chase winners late and dump losers late, and the premiums are the correction of those errors — earned by whoever holds the other side. Lakonishok, Shleifer and Vishny argued this for value in 1994; momentum has almost no other explanation, which is why it sits uneasily in the risk camp.
The consequence is different: a mispricing persists only as long as the mistake does and as long as it is hard to arbitrage. Publish it, pour money into it, and it should shrink. McLean and Pontiff measured that in 2016 across nearly a hundred published anomalies — returns roughly a quarter lower after the sample period ended and more than half lower after publication. Part decay, part something never really there.
Why the choice matters at your desk
Three decisions turn on it. Whether to expect the premium after buying: risk premia persist by construction, mispricings decay as they are learned. How to read a losing stretch: a risk premium losing for three years is doing its job; a mispricing losing for three years may have been arbitraged away. What to hold: a risk you are paid for is one you should take only in the size you can survive — the allocation course's capacity, not tolerance — while a mispricing is a free lunch until the crowd arrives, and the crowding lesson is about the arrival.
Most practitioners settle on a mixed answer: value and size are mostly risk with some behaviour, momentum is mostly behaviour with a large cost of trading it, quality and low volatility are constraints on other investors that may or may not hold. The plain version of that sentence is that nobody knows the split, and a portfolio built on factors should be sized as if the premium might be half what the history says.
In the data
The two stories make one testable prediction apart: the risk story says value shares should fall more in recessions. The value and growth funds through the pandemic are a first look, three dates each: 3 January 2017, 31 August 2020 and 30 December 2022.
From the first date to the second the value fund gained about 18% and the growth fund about 124%, and inside that window the pandemic drawdown was steeper on the value side. One episode is an anecdote; the drawdowns lesson makes it a measurement.
Try it now
- Write two sentences: the risk story for value in your own words, then the behaviour story. Then a third saying which one a value fund losing for three years is evidence for — and notice that it is evidence for both.
- Reopen Why does diversification thin out exactly when it is needed? and write down what it says about assets that fall together in a crisis. The risk story says value is one of them. Keep that line for the drawdowns lesson.
- The index those funds sit inside, over the same window, so the episode has a scale:
Measure 3 January 2017 to 31 August 2020 and compare with the two fund figures above.