Alpha
Alpha is the part of return that market movement does not explain. Positive means the portfolio beat what its beta implied. For most portfolios alpha sits near zero, and over short periods it is barely distinguishable from chance.
- Clearly behind
- The portfolio fell noticeably short of what its market sensitivity implied. Common causes are trading costs, currency effects, or an ill-fitting benchmark.
- Effectively zero
- Where the large majority of portfolios land. Return is explained almost entirely by the market, which is the expected finding for broadly diversified holdings.
- Slightly positive
- A small unexplained surplus. Check the t-statistic: only from a magnitude of about 2 does the value count as statistically sound.
- Unusually high
- Rare over long periods. Usually a factor the model does not know is hiding behind it, such as gold or crypto. A factor model containing those often shrinks the alpha sharply.
How it is calculated
α = R[portfolio] − ( R[risk-free] + β × ( R[benchmark] − R[risk-free] ) )
A portfolio with beta 0.9 returns 9.0 % per year. The index delivers 8.0 %, the risk-free rate 2.0 %. Expected return is therefore 7.4 %. The 1.0 point of excess return over the index is not the alpha.
- Portfolio return
- 9.0 %
- Expected at β 0.9
- 7.4 %
- Excess over index
- 1.0 pp
- Alpha
- 1.6 % p.a.
What the number does not tell you
- An alpha without a t-statistic is a claim. The t-statistic says whether the result differs from chance. Only from a magnitude of about 2 does it count as significant. With less than three years of history, few portfolios reach that.
- It depends entirely on the model. Measured against a single index, any gold or crypto position shows up as alpha. In a factor model that knows those building blocks, the same surplus often all but disappears.
- The benchmark decides too. Measuring a technology portfolio against the MSCI World produces a high alpha in good years and a deeply negative one in bad years. Both describe the sector choice, not skill.
- Costs are often the real effect. Trading fees, spreads and fund costs reliably push alpha into the red. A slightly negative value is the rule rather than an anomaly for active trading.
Related metrics
How Evergrova calculates it
Performance & Risk shows alpha against the chosen index alongside tracking error and information ratio. The factor regression tests the same question more strictly, against market, size, value, momentum, quality, gold, crypto and bonds, and reports the t-statistic.
Calculation: annualised return less the market return explained through beta; in the factor regression, additionally adjusted for known factor premiums.
View the demo portfolioCommon questions
Is alpha the same as excess return?
No. Excess return is simply the difference to the index. Alpha additionally strips out what follows from market sensitivity alone. A portfolio with beta 1.3 has to beat the index substantially just to reach an alpha of zero.
Why is my alpha negative when I am in profit?
Because alpha measures relatively. A 6 % gain against an index at 10 % with a matching beta yields a negative alpha. The figure does not answer whether you earned, but whether more was to be expected.
What is the difference between Jensen alpha and factor alpha?
Jensen alpha measures against a single market index. Factor alpha measures against several known patterns at once and is therefore stricter. What looks like skill against one index often turns out to be a size or value tilt in a factor model.
Last reviewed: 2026-08-08