A fund returned 15% last year. The index returned 10%. Cash returned 4%. The marketing material says “five points ahead of the benchmark” and it is, arithmetically, true.
Whether it means anything depends on a number the marketing material will not lead with: how much market risk the fund took to get there.
The mirage
Beta measures how much an asset moves relative to its benchmark. A beta of 1.5 means that when the index moves one point, this fund has historically moved about one and a half. It is, in effect, a leverage dial on the market itself.
The Capital Asset Pricing Model says an asset should earn the risk-free rate plus beta times the market's excess return. For our fund:
4% + 1.5 × (10% − 4%) = 13%
So of the 15% delivered, 13 points were simply the price of the risk taken. The genuine outperformance — Jensen's alpha — is 2%, not 5%.
Push the beta higher and it disappears entirely. At a beta of 1.83, CAPM expects 14.98%, and a 15% return is alpha of 0.02%. The manager did not pick better companies. They held a more volatile version of the same market and were paid for it, exactly as theory predicts.
| Fund return | Beta | CAPM expects | Jensen's alpha |
|---|---|---|---|
| 15% | 1.00 | 10.00% | +5.00% |
| 15% | 1.50 | 13.00% | +2.00% |
| 15% | 1.83 | 14.98% | +0.02% |
| 20% | 2.00 | 16.00% | +4.00% |
| 11% | 0.70 | 8.20% | +2.80% |
That last row is the one people find counterintuitive. A fund returning 11% produced more alpha than one returning 15%, because it took far less market risk to get there. On a risk-adjusted basis it was the better fund, and in a falling market that difference would have shown up dramatically in the other direction.
Beating the index is not evidence of skill. Beating what your beta predicted is — and only that gap is worth paying a management fee for.
The other half of the trade
High beta is not a trick, and it is not free. The same 1.5 multiplier that turned a 10% market into a 15% return turns a −20% market into roughly −30%. Investors who owned high-beta funds through a long bull run often read their returns as manager skill, then discovered the mechanism in a single quarter.
This is why the shape of the return distribution matters more than its average. Leverage widens both tails.
What alpha does not tell you
Having made the case for alpha, three honest limits on it.
It depends entirely on the benchmark. Measure a small-cap fund against a large-cap index and you will find “alpha” that is really just a size exposure. Multi-factor models exist precisely because much of what single-factor CAPM calls alpha turns out to be systematic exposure to size, value, momentum or quality.
One period proves nothing. Alpha over a single year is dominated by noise. A 2% alpha across three years is comfortably inside the range of chance, and the statistical significance depends on how volatile the excess returns were — not just how large.
It is backward-looking. Past alpha is a weak predictor of future alpha. The broad finding across decades of fund performance research is that very few managers sustain it once fees are deducted.
Treat a positive alpha as a question worth investigating — what did they actually do differently? — rather than as an answer.
Running it on your own holdings
You need four numbers: the fund's return, the benchmark's return over the same period, the risk-free rate for that period, and the beta. The first three are on any factsheet. Beta usually is too, and if you have both price series the calculator will fit it for you by regressing one on the other.
Then ask the only question that matters: after subtracting what the risk alone should have paid, is there anything left?