Decades of academic research found that smaller companies, and companies cheap relative to their fundamentals, historically delivered higher returns than the broad market. Two factors — size and value — with substantial supporting data.
Then came an extended period where they did not. That is the essential context, and any discussion that omits it is selling something.
What changed in 2026
- The debate stayed unresolved. Whether the premium persists, has been arbitraged away, or is simply in a long drawdown remained genuinely contested.
- Factor definitions proliferated. Different providers used materially different measures of "value", making funds less comparable than their labels suggest.
- Implementation cost scrutiny increased. Recognition that trading costs in smaller companies erode a meaningful share of any premium.
- Behavioural risk got more emphasis. The tendency to abandon a tilt after underperformance became the central practical concern.
What the premium is
Two related claims, historically supported:
Size. Smaller companies outperformed larger ones over long periods.
Value. Companies trading cheaply relative to fundamentals outperformed expensive ones.
Combined, small cap value was the strongest historical effect. Explanations divide into two camps, and which you believe determines what you expect going forward.
Risk-based: the premium compensates for genuine additional risk. Small cheap companies are more fragile, more sensitive to downturns, and more likely to fail. Higher expected return is payment for bearing that. If true, the premium should persist, along with the risk that justifies it.
Behavioural: investors systematically overpay for exciting growth companies and underprice dull cheap ones. If true, the premium could shrink as the effect becomes widely known and arbitraged.
The extended recent underperformance is consistent with both — a bad run for a real risk premium, or the effect being competed away.
Underperformance lasts longer than patience
The practical issue, and it matters more than the theoretical debate.
Even taking the historical premium at face value, it did not arrive smoothly. There have been stretches exceeding a decade where a small cap value tilt substantially underperformed the broad market.
A decade is longer than most people's conviction. The realistic sequence is: adopt the tilt, underperform for years, watch a simple index fund do better, conclude the premium is gone, and switch — at which point you have captured the underperformance and none of the recovery.
That behavioural risk is frequently larger than the expected premium. A tilt you will not hold through a bad decade is worse than no tilt at all, and being honest with yourself about that is the actual decision.
| Consideration |
Question to answer |
| Time horizon |
Do I have decades, not years? |
| Tracking error tolerance |
Can I watch the index beat me for a decade? |
| Conviction basis |
Do I understand why I believe this? |
| Implementation cost |
What am I paying to access it? |
| Account type |
Is turnover tax-efficient where I hold it? |
Implementation matters
Accessing the factors costs more than accessing the broad market.
Smaller companies have wider bid-ask spreads, so the fund pays more to trade — see the bid-ask spread. Factor funds also rebalance to maintain their tilt, producing higher turnover than a market-cap index, which adds trading costs and, in a taxable account, distributed gains — see portfolio turnover cost.
Expense ratios are typically higher too.
Together those can consume a meaningful share of any premium, which means a poorly-implemented factor fund may deliver the risk without the return.
Definitions matter as well: providers measure value differently, and two funds labelled the same way can hold substantially different portfolios. Reading what a fund actually screens on is worth the effort.
Common mistakes
- Tilting without understanding the drawdown risk. Then abandoning it at the worst point.
- Comparing against the index over short periods. The horizon is decades.
- Assuming all value funds are equivalent. Definitions vary considerably.
- Ignoring implementation costs. They can consume much of the premium.
- Holding a high-turnover factor fund in a taxable account. Tax drag compounds the cost.
- Concentrating heavily. A tilt is a tilt; a portfolio entirely in one factor is a concentrated bet.
- Treating historical data as a forecast. It is evidence, not a promise.
FAQ
Is the premium gone?
Genuinely unknown. Extended underperformance is consistent with both a real premium in drawdown and an effect that has been arbitraged away. Anyone stating it confidently either way is overreaching.
How much should I tilt?
If at all, modestly — enough to matter, not enough that a decade of underperformance makes the portfolio unrecognisable. Most sensible allocations are a tilt rather than a wholesale replacement.
Is a total market fund missing out?
It holds these companies at market weight, so it participates. A tilt is a deliberate overweight, not access to something otherwise unavailable — see best index funds.
What about other factors?
Momentum, quality, and low volatility have their own literature and their own drawdown histories. The same questions apply to each — see momentum investing and low volatility investing.
Where to go next
For other factor approaches, read momentum investing and low volatility investing. For the broad-market baseline, best index funds.
This is general information, not investment advice. Past performance does not predict future results.