TL;DR
Value spent 2010–2020 trailing the market by 4% per year. Small-cap lagged for three decades 1981–2010. Holding through these stretches is the entire trade — selling at the bottom guarantees you collect the costs without the premium.
The forgotten side of the factor premium
Every factor article shows you the long-run premium. Few of them show you the experience of getting it. Factor investing requires accepting periods of underperformance that would feel absolutely catastrophic in real time. If you don't internalise this upfront, you'll quit at the worst possible moment.
Three well-documented examples:
- Value 2010–2020: pure HML lagged the US market by roughly 4% per year for a decade. Investors who held value funds underperformed cumulatively by about 50% versus the S&P 500. Many sold somewhere between 2017 and 2020 — exactly before the 2020–2024 value rally.
- Size 1981–2010: pure SMB was roughly flat for 30 years. Investors who tilted toward small caps spent three decades being told they were sophisticated while underperforming the broad market.
- Momentum 2009: the factor crashed about 80% in a single quarter as junk stocks rebounded. Investors who panic-sold locked in a loss that recovered fully within 18 months.
These aren't outliers. They're the rule. Factor drawdowns of 5–10 years are normal. 10–20 year drawdowns happen. 30-year drawdowns happen, though rarely.
Why investors quit
The behavioural failure mode is predictable. Three patterns:
- Anchoring to the S&P. The S&P 500 is in every news ticker, every brokerage app, every dinner-party conversation. If you tilt away from it and lag, you experience constant social and emotional pressure. Most investors can't tolerate this for more than 3–5 years.
- Catastrophising the duration. When a factor has lagged for 5 years, investors start asking "what if it's broken forever?" — usually within 6 months of the cyclical low. The narrative shifts from "this is normal" to "this time it's different" exactly when the recovery is about to start.
- Tax-loss harvesting traps. In US taxable accounts especially, the temptation to harvest losses by switching out of a lagging factor fund becomes overwhelming. The replacement is rarely a clean substitute, and the original factor often recovers shortly after.
The mathematical edge of factor investing is roughly 0.5–2% per year. The behavioural cost of poorly-timed exits routinely exceeds 2% per year. The behavioural side determines whether you actually capture the premium.
Tactics that genuinely help
Three pre-commitments that increase the chance of holding through drawdowns:
- Cap your tilt at a level you can actually hold. Most retail investors over-allocate to factor exposure when they're optimistic, then panic-sell when stressed. A 25–40% factor tilt vs total market is usually the upper limit of what's behaviourally tolerable for retail investors. 100% factor exposure looks great on the spreadsheet and disastrous in practice.
- Check performance no more than quarterly. Daily and weekly checks amplify emotional reactions to drawdowns. Quarterly or annual reviews give you enough information without exposing you to short-term noise.
- Pre-commit in writing. Before you start, write down: "I will hold this allocation for at least 10 years regardless of relative performance. Conditions under which I'll change: [specific list]." Specific written commitments survive market stress better than general intentions.
The data supports each of these. AQR's behavioural-finance work shows that retail investors who use systematic rebalancing rules outperform those who make discretionary decisions during stress periods, by roughly 1–2% per year.
How long is "long enough"?
A useful planning baseline: assume any factor tilt can underperform the market for up to 15 years before it starts to compound positively. The shortest periods of underperformance have been 5–7 years; the longest have been over 25 years (size 1981–2010).
For a FIRE planner with a 25-year accumulation, this matters. If you start tilting at 30 and the factor underperforms until 45, you're 60% through your accumulation before the premium starts paying off. You have to hold through the bad part to get the good part.
For investors with shorter horizons (10 years or less), the risk-reward of factor tilts deteriorates sharply. The long-run edge is real but it requires long-run holding to actually capture.
The link to portfolio sizing
The right sizing answer is whatever you'd actually hold through a 10-year underperformance period. For most retail investors, that's:
- 25–40% factor tilt vs total market: comfortable for most
- 40–60% factor tilt: requires explicit behavioural pre-commitment
- 60–100% factor exposure: almost no retail investors can hold through stress without selling
This is the empirically observed truth, not a theoretical limit. The math of factor premiums doesn't change based on your psychology, but your ability to access those premiums absolutely does.
For more on how factor exposure compares to broad-market investing in practice, see our factor investing vs index funds article. Test what a moderate factor tilt does to your FIRE plan in our factor comparison tool — and crucially, look at the worst historical cohorts, not just the median.
Frequently asked questions
- How long can a factor underperform?
- Longer than you'd guess. 10–15 year stretches are well within the historical range, and the size factor underperformed for almost 30 years in the US.
- Should I cut my factor exposure if it's been bad for years?
- Almost certainly not. Selling after a long underperformance period historically locks in the worst possible timing — the next decade has tended to be the best for the factor.
- Is it ever right to abandon a factor entirely?
- Rarely. The only defensible reason is if the underlying mechanism has changed — for example, if size has been arbitraged so thoroughly that its expected premium is genuinely zero. Even then, the more common error is to abandon too early.
Stress-test your own FIRE plan
FIRE Wealth OS runs your savings rate and expenses against every historical market starting point since 1871. Free to use, no card required.