Small-cap stocks outperform large-cap stocks.
Banz · Journal of Financial Economics · published 1981 · original sample 1936–1975 (reported ≈ +5.0%/yr, t ≈ 2.6)
Growth of $1 in the long-short strategy, gross of trading and shorting costs. The dashed line marks the year the paper was published — everything to its right is out-of-sample.
$1 in the long-short strategy, gross of costs. Grey = the sample the paper had; blue = everything since.
The comparison the site is built around.
| Window | Return/yr | Vol | Sharpe | t-stat | Max DD |
|---|---|---|---|---|---|
| Paper's sample (1936–1975) | +2.4% | 11.8% | 0.26 | 0.8 | -44% |
| After publication (1982–2026) | -0.1% | 10.1% | 0.04 | 0.3 | -56% |
| Full sample (1963–2026) | +1.7% | 10.5% | 0.21 | 1.6 | -56% |
Is there a risk-adjusted alpha left once you control for the standard factors? Alpha is annualized; t-stats use Newey–West standard errors.
| Model | Alpha/yr | Alpha t | R² | Biggest loadings |
|---|---|---|---|---|
| CAPM (market only) | +0.9% | 0.7 | 0.07 | Mkt-RF +0.18 |
| Fama–French 3-factor | 0.0% | 0.0 | 1.00 | SMB +1.00, Mkt-RF +0.00, HML +0.00 |
| Fama–French 5-factor | 0.0% | 0.0 | 1.00 | SMB +1.00, Mkt-RF +0.00, HML +0.00 |
| FF5 + momentum * | 0.0% | 0.0 | 1.00 | SMB +1.00, Mkt-RF +0.00, HML +0.00 |
| FF5 + mom, own factor removed | +2.3% | 1.7 | 0.17 | tests for anything left beyond the other five |
* Size (small-minus-big) is itself the SMB factor, so its alpha against a model that contains SMB is zero by construction — the informative row is the one with its own factor removed.
Trailing 3-year Sharpe ratio. Long stretches below zero are common.
| 1960s | +9.6%/yr | Sharpe 0.95 |
| 1970s | +4.7%/yr | Sharpe 0.39 |
| 1980s | -0.4%/yr | Sharpe -0.05 |
| 1990s | -2.1%/yr | Sharpe -0.21 |
| 2000s | +7.1%/yr | Sharpe 0.56 |
| 2010s | -0.4%/yr | Sharpe -0.05 |
| 2020s | -1.6%/yr | Sharpe -0.14 |
Small firms are more exposed to funding, liquidity and recession risk — the premium pays for that. But the risk-based case has always been thin.
Neglect: small caps are under-researched, so mispricing lasts longer.
The clearest decay story of all. The premium essentially vanished right after Banz published — much of the original result was January, delisting bias, and a handful of tiny illiquid names. It survives today mainly as a control variable and as an interaction (small + value, small + quality).
Long-short = the published Fama–French / Carhart factor return. Value-weighted, US equities, monthly, gross of transaction and shorting costs. The original-paper figures are approximate and annualized for comparison.